Installed stretch on a fresh hard drive, using a multiple partition setup, default gnome simple install. Broke my eyeballs and will use up a set of batteries in my mouse weekly to get anything done. Its a single user, single tasking system, with a busted terminal emulator, at best.
Rebooted back to wheezy and printed the "debian trinity repository installation instructions", then rebooted back to stretch to apply them. Did, but got an odd error with the import key line. Fixed trinity.list for apt of course first. But synaptic did run, tde was clear at the bottom of the left column, but I found it and 45 minutes and about 1100MB of downloads later I thought I had enough tde to run. R14.0.5 FWIW.
Rebooted to stretch, was greeted by the trinity login greeter, thought I was home free, but it was quiet for about 45 seconds after I entered the pw. Then boom, I'm looking at the worthless gnome screen again.
During the install, I was asked which manager to use and clicked on tdm-trinity.
But its not sticking, and I can't mount the stretch drive and edit it, theres enough diffs between the ext4's that it refuses to mount from wheezy. And if I add it to wheezy's fstab, the reboot locks up and I have to use the root pw to get a shell and # sign that line back out of wheezy'd fstab. Hopefully I will not have that problem in reverse because there's about 150 GB of stuff on the old wheezy drive to copy over. I did that, and I could read a dir or 3, but when I rebooted to wheezy again, it had to do a 20 minute e2fsck of the wheezy drive, which it did w/o reporting any errors.
Looks like I need help before I really screw this up.
So first, what do I edit or nuke to get rid of the gnome over-ride?
Thanks.
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA512
So first, what do I edit or nuke to get rid of the gnome over-ride?
Thanks.
IMO, since you are installing stretch on a fresh HD, install a minimum environment without any DE at all. Then add TDE. No surprises, you only get TDE :-) Cheers Michele
On Monday 20 August 2018 08:21:43 Michele Calgaro wrote:
So first, what do I edit or nuke to get rid of the gnome over-ride?
Thanks.
IMO, since you are installing stretch on a fresh HD, install a minimum environment without any DE at all. Then add TDE. No surprises, you only get TDE :-) Cheers Michele
Well, I apparently installed the wrong dvd, I was using the first dvd of 9.4 debian, when I should have been using the stretch uspace disc to install linuxcnc. That is after all, the target of this masochistic exercise. :)
So I've now burnt the correct disk and will re-install as if the drive is clean by letting the installer do the partitioning over.
These guys make a pretty good installer simply because they know what it needs and doesn't need to run LinuxCNC right. They don't care if it doesn't run on cousin Molly's washing machine, although with some config effort, it can probably run it, better than the programming it came with.
I certainly had no such problems installing tde on top of the kde I originally loaded onto my Dell running a grizzly G0704 mill on a special wheezy install. So I'll do the same and then install tde on that.
Back later with grins or tears.
Thanks Michelle.
Am Montag, 20. August 2018 schrieb Gene Heskett:
[...] These guys make a pretty good installer simply because they know what it needs and doesn't need to run LinuxCNC right. They don't care if it doesn't run on cousin Molly's washing machine, although with some config effort, it can probably run it, better than the programming it came with.
I certainly had no such problems installing tde on top of the kde I originally loaded onto my Dell running a grizzly G0704 mill on a special wheezy install. So I'll do the same and then install tde on that.
Back later with grins or tears.
Thanks Michelle.
Hi Gene!
You should definitly walk the devuan road. I had a hell of a time with machinekit on BBB, till I found out how to get rid of the systemd stuff. When you run stretch + linuxcnc, you'll most likely get bad surprises (aka. latency > 400000ns) with the 4.*-rtpreempt kernel
Nik
On Monday 20 August 2018 10:41:23 Dr. Nikolaus Klepp wrote:
Am Montag, 20. August 2018 schrieb Gene Heskett:
[...] These guys make a pretty good installer simply because they know what it needs and doesn't need to run LinuxCNC right. They don't care if it doesn't run on cousin Molly's washing machine, although with some config effort, it can probably run it, better than the programming it came with.
I certainly had no such problems installing tde on top of the kde I originally loaded onto my Dell running a grizzly G0704 mill on a special wheezy install. So I'll do the same and then install tde on that.
Back later with grins or tears.
Thanks Michelle.
Hi Gene!
You should definitly walk the devuan road. I had a hell of a time with machinekit on BBB, till I found out how to get rid of the systemd stuff. When you run stretch + linuxcnc, you'll most likely get bad surprises (aka. latency > 400000ns) with the 4.*-rtpreempt kernel
Nik
Its tears so far Nik. There are so many differences just in the ext4's used that one of them should be renamed, they will NOT cross mount, ext4 disk to ext4 mount.
Am Dienstag, 21. August 2018 schrieb Gene Heskett:
On Monday 20 August 2018 10:41:23 Dr. Nikolaus Klepp wrote:
Am Montag, 20. August 2018 schrieb Gene Heskett:
[...] These guys make a pretty good installer simply because they know what it needs and doesn't need to run LinuxCNC right. They don't care if it doesn't run on cousin Molly's washing machine, although with some config effort, it can probably run it, better than the programming it came with.
I certainly had no such problems installing tde on top of the kde I originally loaded onto my Dell running a grizzly G0704 mill on a special wheezy install. So I'll do the same and then install tde on that.
Back later with grins or tears.
Thanks Michelle.
Hi Gene!
You should definitly walk the devuan road. I had a hell of a time with machinekit on BBB, till I found out how to get rid of the systemd stuff. When you run stretch + linuxcnc, you'll most likely get bad surprises (aka. latency > 400000ns) with the 4.*-rtpreempt kernel
Nik
Its tears so far Nik. There are so many differences just in the ext4's used that one of them should be renamed, they will NOT cross mount, ext4 disk to ext4 mount.
Hi Gene!
When you cannot mount the ext4 partions from one another, then there is something very wrong. ext4 can be mounted as ext2 and that should alway work - at least, if the drives and filesystems are ok.
Nik
On Tue, 21 Aug 2018 09:20:43 +0200 "Dr. Nikolaus Klepp" office@klepp.biz wrote:
Its tears so far Nik. There are so many differences just in the ext4's used that one of them should be renamed, they will NOT cross mount, ext4 disk to ext4 mount.
Hi Gene!
When you cannot mount the ext4 partions from one another, then there is something very wrong. ext4 can be mounted as ext2 and that should alway work - at least, if the drives and filesystems are ok.
No, you are wrong. There were options added to ext4 which made it incompatible with older kernels. And a few years ago they made this options default. Debian wheezy kernel (3.2) would not mount ext4 created in debian stretch.
Am Dienstag, 21. August 2018 schrieb Nick Koretsky:
On Tue, 21 Aug 2018 09:20:43 +0200 "Dr. Nikolaus Klepp" office@klepp.biz wrote:
Its tears so far Nik. There are so many differences just in the ext4's used that one of them should be renamed, they will NOT cross mount, ext4 disk to ext4 mount.
Hi Gene!
When you cannot mount the ext4 partions from one another, then there is something very wrong. ext4 can be mounted as ext2 and that should alway work - at least, if the drives and filesystems are ok.
No, you are wrong. There were options added to ext4 which made it incompatible with older kernels. And a few years ago they made this options default. Debian wheezy kernel (3.2) would not mount ext4 created in debian stretch.
Hi!
Now that's a gotcha I did not know. How did it come that this was thought of beeing a good idea?
Nik
On Tue, 21 Aug 2018 12:55:28 +0200 "Dr. Nikolaus Klepp" office@klepp.biz wrote:
Am Dienstag, 21. August 2018 schrieb Nick Koretsky:
On Tue, 21 Aug 2018 09:20:43 +0200 "Dr. Nikolaus Klepp" office@klepp.biz wrote:
Its tears so far Nik. There are so many differences just in the ext4's used that one of them should be renamed, they will NOT cross mount, ext4 disk to ext4 mount.
Hi Gene!
When you cannot mount the ext4 partions from one another, then there is something very wrong. ext4 can be mounted as ext2 and that should alway work - at least, if the drives and filesystems are ok.
No, you are wrong. There were options added to ext4 which made it incompatible with older kernels. And a few years ago they made this options default. Debian wheezy kernel (3.2) would not mount ext4 created in debian stretch.
Hi!
Now that's a gotcha I did not know. How did it come that this was thought of beeing a good idea?
Nik
Yep. Exactly my thought when i spent 3-hours with a server refusing to boot after migration to a new hdd (i used a stretch flash to copy). Who the fuck toughs it was a good idea?!! Why not call it ext5 or ext4a or whatever?!!
On Tuesday 21 August 2018 08:21:12 Nick Koretsky wrote:
On Tue, 21 Aug 2018 12:55:28 +0200
"Dr. Nikolaus Klepp" office@klepp.biz wrote:
Am Dienstag, 21. August 2018 schrieb Nick Koretsky:
On Tue, 21 Aug 2018 09:20:43 +0200
"Dr. Nikolaus Klepp" office@klepp.biz wrote:
Its tears so far Nik. There are so many differences just in the ext4's used that one of them should be renamed, they will NOT cross mount, ext4 disk to ext4 mount.
Hi Gene!
When you cannot mount the ext4 partions from one another, then there is something very wrong. ext4 can be mounted as ext2 and that should alway work - at least, if the drives and filesystems are ok.
No, you are wrong. There were options added to ext4 which made it incompatible with older kernels. And a few years ago they made this options default. Debian wheezy kernel (3.2) would not mount ext4 created in debian stretch.
Hi!
Now that's a gotcha I did not know. How did it come that this was thought of beeing a good idea?
Nik
Yep. Exactly my thought when i spent 3-hours with a server refusing to boot after migration to a new hdd (i used a stretch flash to copy). Who the fuck toughs it was a good idea?!! Why not call it ext5 or ext4a or whatever?!!
+100
Now the question of the day is how the hell do we fix it?
On Tue, 21 Aug 2018 09:08:55 -0400 Gene Heskett gheskett@shentel.net wrote:
On Tuesday 21 August 2018 08:21:12 Nick Koretsky wrote:
On Tue, 21 Aug 2018 12:55:28 +0200
"Dr. Nikolaus Klepp" office@klepp.biz wrote:
Am Dienstag, 21. August 2018 schrieb Nick Koretsky:
On Tue, 21 Aug 2018 09:20:43 +0200
"Dr. Nikolaus Klepp" office@klepp.biz wrote:
Its tears so far Nik. There are so many differences just in the ext4's used that one of them should be renamed, they will NOT cross mount, ext4 disk to ext4 mount.
Hi Gene!
When you cannot mount the ext4 partions from one another, then there is something very wrong. ext4 can be mounted as ext2 and that should alway work - at least, if the drives and filesystems are ok.
No, you are wrong. There were options added to ext4 which made it incompatible with older kernels. And a few years ago they made this options default. Debian wheezy kernel (3.2) would not mount ext4 created in debian stretch.
Hi!
Now that's a gotcha I did not know. How did it come that this was thought of beeing a good idea?
Nik
Yep. Exactly my thought when i spent 3-hours with a server refusing to boot after migration to a new hdd (i used a stretch flash to copy). Who the fuck toughs it was a good idea?!! Why not call it ext5 or ext4a or whatever?!!
+100
Now the question of the day is how the hell do we fix it?
If you need a shared partition between wheezy and stretch create it in wheezy.
Dne út 21. srpna 2018 Nick Koretsky napsal(a):
On Tue, 21 Aug 2018 09:08:55 -0400
Gene Heskett gheskett@shentel.net wrote:
On Tuesday 21 August 2018 08:21:12 Nick Koretsky wrote:
On Tue, 21 Aug 2018 12:55:28 +0200
"Dr. Nikolaus Klepp" office@klepp.biz wrote:
Am Dienstag, 21. August 2018 schrieb Nick Koretsky:
On Tue, 21 Aug 2018 09:20:43 +0200
"Dr. Nikolaus Klepp" office@klepp.biz wrote:
> Its tears so far Nik. There are so many differences just in > the ext4's used that one of them should be renamed, they > will NOT cross mount, ext4 disk to ext4 mount.
Hi Gene!
When you cannot mount the ext4 partions from one another, then there is something very wrong. ext4 can be mounted as ext2 and that should alway work - at least, if the drives and filesystems are ok.
No, you are wrong. There were options added to ext4 which made it incompatible with older kernels. And a few years ago they made this options default. Debian wheezy kernel (3.2) would not mount ext4 created in debian stretch.
Hi!
Now that's a gotcha I did not know. How did it come that this was thought of beeing a good idea?
Nik
Yep. Exactly my thought when i spent 3-hours with a server refusing to boot after migration to a new hdd (i used a stretch flash to copy). Who the fuck toughs it was a good idea?!! Why not call it ext5 or ext4a or whatever?!!
+100
Now the question of the day is how the hell do we fix it?
If you need a shared partition between wheezy and stretch create it in wheezy.
I recommend upgrading to wheezy-backports kernel.
Cheers
On Tuesday 21 August 2018 09:47:19 Slávek Banko wrote:
Dne út 21. srpna 2018 Nick Koretsky napsal(a):
On Tue, 21 Aug 2018 09:08:55 -0400
Gene Heskett gheskett@shentel.net wrote:
On Tuesday 21 August 2018 08:21:12 Nick Koretsky wrote:
On Tue, 21 Aug 2018 12:55:28 +0200
"Dr. Nikolaus Klepp" office@klepp.biz wrote:
Am Dienstag, 21. August 2018 schrieb Nick Koretsky:
On Tue, 21 Aug 2018 09:20:43 +0200
"Dr. Nikolaus Klepp" office@klepp.biz wrote: > > Its tears so far Nik. There are so many differences just > > in the ext4's used that one of them should be renamed, > > they will NOT cross mount, ext4 disk to ext4 mount. > > Hi Gene! > > When you cannot mount the ext4 partions from one another, > then there is something very wrong. ext4 can be mounted as > ext2 and that should alway work - at least, if the drives > and filesystems are ok.
No, you are wrong. There were options added to ext4 which made it incompatible with older kernels. And a few years ago they made this options default. Debian wheezy kernel (3.2) would not mount ext4 created in debian stretch.
Hi!
Now that's a gotcha I did not know. How did it come that this was thought of beeing a good idea?
Nik
Yep. Exactly my thought when i spent 3-hours with a server refusing to boot after migration to a new hdd (i used a stretch flash to copy). Who the fuck toughs it was a good idea?!! Why not call it ext5 or ext4a or whatever?!!
+100
Now the question of the day is how the hell do we fix it?
If you need a shared partition between wheezy and stretch create it in wheezy.
I recommend upgrading to wheezy-backports kernel.
Cheers
From the plethora of errors I'm getting in a synaptic refresh, the wheezy repos seem to have been shut down. Support ended June 30 IIRC. Even backports is 404. The rest are address not found.
On Tue, 21 Aug 2018 10:10:56 -0400 Gene Heskett gheskett@shentel.net wrote:
From the plethora of errors I'm getting in a synaptic refresh, the wheezy repos seem to have been shut down. Support ended June 30 IIRC. Even backports is 404. The rest are address not found.
Hm? http://ftp.debian.org/ still have wheezy. The debian repos are never gone, the older one just get moved to http://archive.debian.org
On Tuesday 21 August 2018 10:57:02 Nick Koretsky wrote:
On Tue, 21 Aug 2018 10:10:56 -0400
Gene Heskett gheskett@shentel.net wrote:
From the plethora of errors I'm getting in a synaptic refresh, the wheezy repos seem to have been shut down. Support ended June 30 IIRC. Even backports is 404. The rest are address not found.
Hm? http://ftp.debian.org/ still have wheezy.
but no key files, something aptitude doesn't tolerate, and synaptic complains about at length. or just plain gone. If synaptic will copy/paste: Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/Release.gpg Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/main/binary-i386/Pac... Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/contrib/binary-i386/... Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/non-free/binary-i386... Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/contrib/i18n/Transla... Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/contrib/i18n/Transla... Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/main/i18n/Translatio... Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/main/i18n/Translatio... Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/non-free/i18n/Transl... Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/non-free/i18n/Transl... Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://mozilla.debian.net/dists/wheezy-backports/firefox-release/binary-i386... 404 Not Found [IP: 149.20.4.15 80] Some index files failed to download. They have been ignored, or old ones used instead.
The debian repos are never gone, the older one just get moved to http://archive.debian.org
So I should edit the repos to be archive.debian.org? Is there just one, no mirrors?
Thanks Nick.
On Tue, 21 Aug 2018 13:18:38 -0400 Gene Heskett gheskett@shentel.net wrote:
On Tue, 21 Aug 2018 10:10:56 -0400
Gene Heskett gheskett@shentel.net wrote:
From the plethora of errors I'm getting in a synaptic refresh, the wheezy repos seem to have been shut down. Support ended June 30 IIRC. Even backports is 404. The rest are address not found.
Hm? http://ftp.debian.org/ still have wheezy.
but no key files, something aptitude doesn't tolerate, and synaptic complains about at length. or just plain gone. If synaptic will copy/paste: Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/Release.gpg Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/main/binary-i386/Pac... Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/contrib/binary-i386/... Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/non-free/binary-i386... Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/contrib/i18n/Transla... Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/contrib/i18n/Transla... Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/main/i18n/Translatio... Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/main/i18n/Translatio... Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/non-free/i18n/Transl... Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/non-free/i18n/Transl... Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://mozilla.debian.net/dists/wheezy-backports/firefox-release/binary-i386... 404 Not Found [IP: 149.20.4.15 80] Some index files failed to download. They have been ignored, or old ones used instead.
The us.debian.net mirror is down apparently. Use a different one. I have a wheezy machine and just ran a apt-get update without problem.
The debian repos are never gone, the older one just get moved to http://archive.debian.org
So I should edit the repos to be archive.debian.org? Is there just one, no mirrors?
No, wheezy is still in the main repos, it have not yet moved to archive.
On Tuesday 21 August 2018 13:43:18 Nick Koretsky wrote:
On Tue, 21 Aug 2018 13:18:38 -0400
Gene Heskett gheskett@shentel.net wrote:
On Tue, 21 Aug 2018 10:10:56 -0400
Gene Heskett gheskett@shentel.net wrote:
From the plethora of errors I'm getting in a synaptic refresh, the wheezy repos seem to have been shut down. Support ended June 30 IIRC. Even backports is 404. The rest are address not found.
Hm? http://ftp.debian.org/ still have wheezy.
but no key files, something aptitude doesn't tolerate, and synaptic complains about at length. or just plain gone. If synaptic will copy/paste: Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/Release.gpg Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/main/binary- i386/Packages Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/contrib/bina ry-i386/Packages Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/non-free/bin ary-i386/Packages Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/contrib/i18n /Translation-en_US Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/contrib/i18n /Translation-en Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/main/i18n/Tr anslation-en_US Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/main/i18n/Tr anslation-en Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/non-free/i18 n/Translation-en_US Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/non-free/i18 n/Translation-en Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://mozilla.debian.net/dists/wheezy-backports/firefox-release/bin ary-i386/Packages 404 Not Found [IP: 149.20.4.15 80] Some index files failed to download. They have been ignored, or old ones used instead.
The us.debian.net mirror is down apparently. Use a different one. I have a wheezy machine and just ran a apt-get update without problem.
The debian repos are never gone, the older one just get moved to http://archive.debian.org
So I should edit the repos to be archive.debian.org? Is there just one, no mirrors?
No, wheezy is still in the main repos, it have not yet moved to archive.
I took the ".us" out of the hostnames, now my fail list is 2x longer, no hostname associated failures. If they still exist, I need the full url's. please?
On 08/21/2018 11:47 AM, Gene Heskett wrote:
I took the ".us" out of the hostnames, now my fail list is 2x longer, no hostname associated failures. If they still exist, I need the full url's. please?
Try: http://ftp.debian.org/debian/dists/wheezy/
This page should have whatever else you need to know (although it says Wheezy has already been moved to the archive, when in fact it hasn't): https://wiki.debian.org/DebianWheezy
On 08/21/2018 12:10 PM, Dan Youngquist wrote:
Actually I guess that would be: http://ftp.debian.org/debian/dists/
Then you have wheezy, wheezy-updates, and wheezy-backports.
On Tuesday 21 August 2018 15:10:00 Dan Youngquist wrote:
On 08/21/2018 11:47 AM, Gene Heskett wrote:
I took the ".us" out of the hostnames, now my fail list is 2x longer, no hostname associated failures. If they still exist, I need the full url's. please?
Thanks Dan, it works, but nothing here I don't already have.
This page should have whatever else you need to know (although it says Wheezy has already been moved to the archive, when in fact it hasn't): https://wiki.debian.org/DebianWheezy
To unsubscribe, e-mail: trinity-users-unsubscribe@lists.pearsoncomputing.net For additional commands, e-mail: trinity-users-help@lists.pearsoncomputing.net Read list messages on the web archive: http://trinity-users.pearsoncomputing.net/ Please remember not to top-post: http://trinity.pearsoncomputing.net/mailing_lists/#top-posting
On Tuesday 21 August 2018 15:42:09 Gene Heskett wrote:
On Tuesday 21 August 2018 15:10:00 Dan Youngquist wrote:
On 08/21/2018 11:47 AM, Gene Heskett wrote:
I took the ".us" out of the hostnames, now my fail list is 2x longer, no hostname associated failures. If they still exist, I need the full url's. please?
Thanks Dan, it works, but nothing here I don't already have.
This page should have whatever else you need to know (although it says Wheezy has already been moved to the archive, when in fact it hasn't): https://wiki.debian.org/DebianWheezy
Now, I just did 2 more installs, turning off my speakers first, neither of which installed a byte of grub to the /boot partition. So I rebooted a third time to the live image, and found no trace of a grub in any dir of the live image boot. So there is no chance of fixing the installers hiccups. I think this last time I had it save the logs to the boot partition on this drive, but I'll be damned if I can find them.
I did find a way around formatting the disk, but the first time it mentioned installing grub it wanted to put it in the MBR of /dev/sda1, which I refused as I'll goto the bios boot menu to boot from the chosen disk. So far I have managed to keep it from screwing up the boot from sda.
WTF? I am running out of patience, and didn't have a whole lot to start with. And at 83, I don't have enough hair to keep from sunburning my head if I'm out for half an hour.
Next, i'll do a full shutdown and exchange the drives, putting the new one in the sda position, and this one in the sdb position (I have a 3 drive quick change cage from tiger direct in this box) and see if it will boot this system from what will become /dev/sdb.
So here goes nothing...
- To unsubscribe, e-mail:
trinity-users-unsubscribe@lists.pearsoncomputing.net For additional commands, e-mail: trinity-users-help@lists.pearsoncomputing.net Read list messages on the web archive: http://trinity-users.pearsoncomputing.net/ Please remember not to top-post: http://trinity.pearsoncomputing.net/mailing_lists/#top-posting
On Tue August 21 2018 14:43:38 Gene Heskett wrote:
Now, I just did 2 more installs, turning off my speakers first, neither of which installed a byte of grub to the /boot partition. So I rebooted a third time to the live image, and found no trace of a grub in any dir of the live image boot. So there is no chance of fixing the installers hiccups. I think this last time I had it save the logs to the boot partition on this drive, but I'll be damned if I can find them.
I did find a way around formatting the disk, but the first time it mentioned installing grub it wanted to put it in the MBR of /dev/sda1, which I refused as I'll goto the bios boot menu to boot from the chosen disk. So far I have managed to keep it from screwing up the boot from sda.
Hi Gene,
Is there some Trinity bug being reported here that I am overlooking?
Or would this thread be better directed to a Debian or Devuan mailing list related to whichever installer you are experiencing problems with, and where the list members might be more able to offer relevant advice and solutions?
--Mike
Mike Bird composed on 2018-08-21 15:42 (UTC-0700):
Is there some Trinity bug being reported here that I am overlooking?
No Trinity bug. Octogenarian foibles coupled with cryptic debian-installer.
Or would this thread be better directed to a Debian or Devuan mailing list related to whichever installer you are experiencing problems with, and where the list members might be more able to offer relevant advice and solutions?
If you were a debian-user@lists.debian.org subscriber you might know he did that, but got little other than complaints about Gene's verbosity and failing to follow directions, or unhelpful "attempts" to help. It didn't help that his subject line says nothing about his problem(s) on a list where most would-be helpers are very picky people who apparently trash such posts without even skimming their content.
Not that this thread's $SUBJECT is likely to be much of a draw, but at least here people aren't so selective about getting involved.
On Tuesday 21 August 2018 20:34:14 Felix Miata wrote:
Mike Bird composed on 2018-08-21 15:42 (UTC-0700):
Is there some Trinity bug being reported here that I am overlooking?
No Trinity bug. Octogenarian foibles coupled with cryptic debian-installer.
Or would this thread be better directed to a Debian or Devuan mailing list related to whichever installer you are experiencing problems with, and where the list members might be more able to offer relevant advice and solutions?
If you were a debian-user@lists.debian.org subscriber you might know he did that, but got little other than complaints about Gene's verbosity and failing to follow directions, or unhelpful "attempts" to help. It didn't help that his subject line says nothing about his problem(s) on a list where most would-be helpers are very picky people who apparently trash such posts without even skimming their content.
Not that this thread's $SUBJECT is likely to be much of a draw, but at least here people aren't so selective about getting involved.
Thanks Felix.
On Tuesday 21 August 2018 18:42:26 Mike Bird wrote:
On Tue August 21 2018 14:43:38 Gene Heskett wrote:
Now, I just did 2 more installs, turning off my speakers first, neither of which installed a byte of grub to the /boot partition. So I rebooted a third time to the live image, and found no trace of a grub in any dir of the live image boot. So there is no chance of fixing the installers hiccups. I think this last time I had it save the logs to the boot partition on this drive, but I'll be damned if I can find them.
I did find a way around formatting the disk, but the first time it mentioned installing grub it wanted to put it in the MBR of /dev/sda1, which I refused as I'll goto the bios boot menu to boot from the chosen disk. So far I have managed to keep it from screwing up the boot from sda.
Hi Gene,
Is there some Trinity bug being reported here that I am overlooking?
Or would this thread be better directed to a Debian or Devuan mailing list related to whichever installer you are experiencing problems with, and where the list members might be more able to offer relevant advice and solutions?
--Mike
Maybe Mike, but I've gotten way more help here than on the emc or debian-user lists.
Now, you might want a trinity bug report just to keep this a wee bit closer to the list topic:
The only problem I have with trinity is I can't install it anywhere because the keyserver will not give me a key from the key retrieval command line shown in the "Debian Trinity Repository Installation Instructions", printed out from the wiki page of that title, its refusing my connection but I can ping keyserver.quickbuild.io just fine.
Once I have that, and this next install actually installs a grub2 booter, I ought to be off to the races. The rest of it should be just a-h and elbows efforts until I have stretch running as smooth as wheezy is right now. As I've probably forgotten 50% of what I did, I know some of it will probably have to be re-invented, but I'll at least see the light at the end of the tunnel.
On Tue August 21 2018 17:48:42 Gene Heskett wrote:
The only problem I have with trinity is I can't install it anywhere because the keyserver will not give me a key from the key retrieval command line shown in the "Debian Trinity Repository Installation Instructions", printed out from the wiki page of that title, its refusing my connection but I can ping keyserver.quickbuild.io just fine.
apt-key adv --keyserver keyserver.quickbuild.io --recv-keys F5CFC95C
... works for me. What error do you get? Do you have dirmngr installed?
--Mike
On Tuesday 21 August 2018 21:01:23 Mike Bird wrote:
On Tue August 21 2018 17:48:42 Gene Heskett wrote:
The only problem I have with trinity is I can't install it anywhere because the keyserver will not give me a key from the key retrieval command line shown in the "Debian Trinity Repository Installation Instructions", printed out from the wiki page of that title, its refusing my connection but I can ping keyserver.quickbuild.io just fine.
apt-key adv --keyserver keyserver.quickbuild.io --recv-keys F5CFC95C
... works for me. What error do you get? Do you have dirmngr installed?
--Mike
Yes. The server is refusing my connections. Is this what I might get if ssh/sshd isn't running? I have to admit I haven't checked.
On Tue August 21 2018 19:38:37 Gene Heskett wrote:
On Tuesday 21 August 2018 21:01:23 Mike Bird wrote:
On Tue August 21 2018 17:48:42 Gene Heskett wrote:
The only problem I have with trinity is I can't install it anywhere because the keyserver will not give me a key from the key retrieval command line shown in the "Debian Trinity Repository Installation Instructions", printed out from the wiki page of that title, its refusing my connection but I can ping keyserver.quickbuild.io just fine.
apt-key adv --keyserver keyserver.quickbuild.io --recv-keys F5CFC95C
... works for me. What error do you get? Do you have dirmngr installed?
--Mike
Yes. The server is refusing my connections. Is this what I might get if ssh/sshd isn't running? I have to admit I haven't checked.
What error do you get, Gene?
Do you have a firewall which might be blocking TCP traffic on port 11371?
What exactly happens if you "telnet keyserver.quickbuild.io 11371" and then hit return a few times?
--Mike
On Wednesday 22 August 2018 00:31:13 Mike Bird wrote:
On Tue August 21 2018 19:38:37 Gene Heskett wrote:
On Tuesday 21 August 2018 21:01:23 Mike Bird wrote:
On Tue August 21 2018 17:48:42 Gene Heskett wrote:
The only problem I have with trinity is I can't install it anywhere because the keyserver will not give me a key from the key retrieval command line shown in the "Debian Trinity Repository Installation Instructions", printed out from the wiki page of that title, its refusing my connection but I can ping keyserver.quickbuild.io just fine.
apt-key adv --keyserver keyserver.quickbuild.io --recv-keys F5CFC95C
... works for me. What error do you get? Do you have dirmngr installed?
--Mike
Yes. The server is refusing my connections. Is this what I might get if ssh/sshd isn't running? I have to admit I haven't checked.
What error do you get, Gene?
As I've repeated several time Mike, connection refused. Thats it. All of it.
Do you have a firewall which might be blocking TCP traffic on port 11371?
What exactly happens if you "telnet keyserver.quickbuild.io 11371" and then hit return a few times?
gene@coyote:/$ telnet keyserver.quickbuild.io 11371 Trying 192.119.205.242... Connected to keyserver.quickbuild.io. Escape character is '^]'.
HTTP/1.1 400 Bad Request
Connection closed by foreign host. After 2 returns.
--Mike
To unsubscribe, e-mail: trinity-users-unsubscribe@lists.pearsoncomputing.net For additional commands, e-mail: trinity-users-help@lists.pearsoncomputing.net Read list messages on the web archive: http://trinity-users.pearsoncomputing.net/ Please remember not to top-post: http://trinity.pearsoncomputing.net/mailing_lists/#top-posting
Dne st 22. srpna 2018 Gene Heskett napsal(a):
On Wednesday 22 August 2018 00:31:13 Mike Bird wrote:
On Tue August 21 2018 19:38:37 Gene Heskett wrote:
On Tuesday 21 August 2018 21:01:23 Mike Bird wrote:
On Tue August 21 2018 17:48:42 Gene Heskett wrote:
The only problem I have with trinity is I can't install it anywhere because the keyserver will not give me a key from the key retrieval command line shown in the "Debian Trinity Repository Installation Instructions", printed out from the wiki page of that title, its refusing my connection but I can ping keyserver.quickbuild.io just fine.
apt-key adv --keyserver keyserver.quickbuild.io --recv-keys F5CFC95C
... works for me. What error do you get? Do you have dirmngr installed?
--Mike
Yes. The server is refusing my connections. Is this what I might get if ssh/sshd isn't running? I have to admit I haven't checked.
What error do you get, Gene?
As I've repeated several time Mike, connection refused. Thats it. All of it.
Do you have a firewall which might be blocking TCP traffic on port 11371?
What exactly happens if you "telnet keyserver.quickbuild.io 11371" and then hit return a few times?
gene@coyote:/$ telnet keyserver.quickbuild.io 11371 Trying 192.119.205.242... Connected to keyserver.quickbuild.io. Escape character is '^]'.
HTTP/1.1 400 Bad Request
Connection closed by foreign host. After 2 returns.
--Mike
Instead of getting a key from the server, you can manually download and install the deb package. You choose the appropriate file with the name *_all.deb:
http://mirror.ppa.trinitydesktop.org/trinity/trinity-r14.0.0/debian/pool/mai...
Cheers
On Wed August 22 2018 05:07:00 Gene Heskett wrote:
gene@coyote:/$ telnet keyserver.quickbuild.io 11371 Trying 192.119.205.242... Connected to keyserver.quickbuild.io. Escape character is '^]'.
HTTP/1.1 400 Bad Request
Connection closed by foreign host. After 2 returns.
Hi Gene,
That's the keyserver accepting your connection. So let's get the key you need. This may wrap but it needs to be typed all on one line:
wget -O keyfile 'http://keyserver.quickbuild.io:11371/pks/lookup?op=get&search=0xF5CFC95C'
Does that work? You should have now have the key you want in "keyfile" (wrapped in a little HTML which you can easily remove).
--Mike
On Wednesday 22 August 2018 10:31:32 Mike Bird wrote:
On Wed August 22 2018 05:07:00 Gene Heskett wrote:
gene@coyote:/$ telnet keyserver.quickbuild.io 11371 Trying 192.119.205.242... Connected to keyserver.quickbuild.io. Escape character is '^]'.
HTTP/1.1 400 Bad Request
Connection closed by foreign host. After 2 returns.
Hi Gene,
That's the keyserver accepting your connection. So let's get the key you need. This may wrap but it needs to be typed all on one line:
The 64$ question, is why doesn't the normal command from the wiki page work?
wget -O keyfile 'http://keyserver.quickbuild.io:11371/pks/lookup?op=get&search=0xF5CFC 95C'
I'll print this so I can do it when the next install is done.
But it doesn't work from this wheezy install either. Just tried it, got a rec from wget to read the man page. turned off the word-wrap, here is the line I tried.
gene@coyote:/$ wget -O keyfile 'http://keyserver.quickbuild.io:11371/pks/lookup?op=get&search=0xF5CFC95C Cannot specify both -k and -O if multiple URLs are given, or in combination with -p or -r. See the manual for details.
Usage: wget [OPTION]... [URL]...
wget seems to be broken here. But it worked for something a week ago...
Does that work? You should have now have the key you want in "keyfile" (wrapped in a little HTML which you can easily remove).
--Mike
On Wed August 22 2018 07:59:49 Gene Heskett wrote:
gene@coyote:/$ wget -O keyfile 'http://keyserver.quickbuild.io:11371/pks/lookup?op=get&search=0xF5CFC95C Cannot specify both -k and -O if multiple URLs are given, or in combination with -p or -r. See the manual for details.
Usage: wget [OPTION]... [URL]...
wget seems to be broken here. But it worked for something a week ago...
If you enter the command "alias" is there anything related to wget in the resulting output? If so, what?
Also, are you using the right wget?
$ which wget /usr/bin/wget
And if you have debsums:
$ debsums wget | grep bin /usr/bin/wget (lots of spaces) OK
--Mike
Am Mittwoch 22 August 2018 schrieb Gene Heskett:
gene@coyote:/$ wget -O keyfile 'http://keyserver.quickbuild.io:11371/pks/lookup?op=get&search=0xF5CFC95C
Hi Gene,
looks like you forgot the ' at the end of your command...
HTH
Kind Regards, Stefan
On Thursday 23 August 2018 07:21:47 Stefan Krusche wrote:
Am Mittwoch 22 August 2018 schrieb Gene Heskett:
gene@coyote:/$ wget -O keyfile 'http://keyserver.quickbuild.io:11371/pks/lookup?op=get&search=0xF5C FC95C
Hi Gene,
looks like you forgot the ' at the end of your command...
That was probably a wet ram memory error, it was there. However, the example in the wiki page has no quotes and it is also "connection refused". This is the:
sudo apt-get adv --keyserver keyserver.quickbuild.io --recv-key F5CFC95C
line from the wiki page, is that page in error? If the above line is correct, and the server does not need rebooting, what am I doing wrong?, and if the above line is incorrect, please fix the wiki.
Thats been my whole point all along.
HTH
Kind Regards, Stefan
To unsubscribe, e-mail: trinity-users-unsubscribe@lists.pearsoncomputing.net For additional commands, e-mail: trinity-users-help@lists.pearsoncomputing.net Read list messages on the web archive: http://trinity-users.pearsoncomputing.net/ Please remember not to top-post: http://trinity.pearsoncomputing.net/mailing_lists/#top-posting
On 08/23/2018 07:01 AM, Gene Heskett wrote:
sudo apt-get adv --keyserver keyserver.quickbuild.io --recv-key F5CFC95C
That should be apt-key, not apt-get. Is that typo only here, or did you make it when you tried the command also?
On Thursday 23 August 2018 10:39:09 Dan Youngquist wrote:
On 08/23/2018 07:01 AM, Gene Heskett wrote:
sudo apt-get adv --keyserver keyserver.quickbuild.io --recv-key F5CFC95C
That should be apt-key, not apt-get. Is that typo only here, or did you make it when you tried the command also?
I am pretty sure it was apt-key, but I'll do another install and dbl-check. Muscle memory, mumble mumble.
On Thursday 23 August 2018 07:21:47 Stefan Krusche wrote:
Am Mittwoch 22 August 2018 schrieb Gene Heskett:
gene@coyote:/$ wget -O keyfile 'http://keyserver.quickbuild.io:11371/pks/lookup?op=get&search=0xF5C FC95C
Hi Gene,
looks like you forgot the ' at the end of your command...
HTH
Kind Regards, Stefan
See my other reply Stefan, to which I'll add that I will not attempt another stretch install since its at least an hours work, until I have been assured that getting the trinity key does work, preferably as advertised in the wiki.
Am Donnerstag 23 August 2018 schrieb Gene Heskett:
See my other reply Stefan, to which I'll add that I will not attempt another stretch install since its at least an hours work, until I have been assured that getting the trinity key does work, preferably as advertised in the wiki.
I'm sorry it didn't help. I was just a quick guess.
Kind regards, Stefan
On Thursday 23 August 2018 15:59:49 Stefan Krusche wrote:
Am Donnerstag 23 August 2018 schrieb Gene Heskett:
See my other reply Stefan, to which I'll add that I will not attempt another stretch install since its at least an hours work, until I have been assured that getting the trinity key does work, preferably as advertised in the wiki.
I'm sorry it didn't help. I was just a quick guess.
Kind regards, Stefan
Don't be sorry Stefan, at least you tried. And I thank you for trying.
To unsubscribe, e-mail: trinity-users-unsubscribe@lists.pearsoncomputing.net For additional commands, e-mail: trinity-users-help@lists.pearsoncomputing.net Read list messages on the web archive: http://trinity-users.pearsoncomputing.net/ Please remember not to top-post: http://trinity.pearsoncomputing.net/mailing_lists/#top-posting
On Thursday 23 August 2018 16:39:49 Gene Heskett wrote:
On Thursday 23 August 2018 15:59:49 Stefan Krusche wrote:
Am Donnerstag 23 August 2018 schrieb Gene Heskett:
See my other reply Stefan, to which I'll add that I will not attempt another stretch install since its at least an hours work, until I have been assured that getting the trinity key does work, preferably as advertised in the wiki.
I'm sorry it didn't help. I was just a quick guess.
Kind regards, Stefan
Don't be sorry Stefan, at least you tried. And I thank you for trying.
And after about 3 more installs, I got it to work, but in my calculations for partition size, I forgot theres around 80 GB in /var, my web pages etc, which was going to be in the / partition, so 40Gb was soon filled up in copying it across. But lots of empty space left in home, so I fired up gparted to shrink the top of home and make room for a new /opt partition. But gparted refuses to do that resize/move. So it looks like another whole new install tomorrow, except tomorrow will be after a new quick change drive cage arrives
Damn this is getting old...
I finally hauled it out to the front deck within reach of an air hose and gave it its annual D&C, plus replaced the A slot cable (it wasn't red either) and put 3 layers of scotch tape on the back of the slot doors spring to make sure the drive was firmly seated as there were times it didn't register in the bios. Now that seems to be good, but I trolled Newegg and found a 4 by 3 drive cage, which I needed a decade back when I built this box from scratch. So thats on order, fedex=Monday here I'd imagine. TDE installed just like it was supposed to.
So I'm going to leave it be till that bit of rebuilding is done.
Many many thanks to those that helped or tried to help, Its appreciated, a lot.
On Thu August 23 2018 19:52:22 Gene Heskett wrote:
And after about 3 more installs, I got it to work, but in my calculations for partition size, I forgot theres around 80 GB in /var, my web pages etc, which was going to be in the / partition, so 40Gb was soon filled up in copying it across. But lots of empty space left in home, so I fired up gparted to shrink the top of home and make room for a new /opt partition. But gparted refuses to do that resize/move. So it looks like another whole new install tomorrow, except tomorrow will be after a new quick change drive cage arrives
Some possibilities you could consider are (1) fewer partitions and (2) LVM.
--Mike
On Thursday 23 August 2018 23:24:37 Mike Bird wrote:
On Thu August 23 2018 19:52:22 Gene Heskett wrote:
And after about 3 more installs, I got it to work, but in my calculations for partition size, I forgot theres around 80 GB in /var, my web pages etc, which was going to be in the / partition, so 40Gb was soon filled up in copying it across. But lots of empty space left in home, so I fired up gparted to shrink the top of home and make room for a new /opt partition. But gparted refuses to do that resize/move. So it looks like another whole new install tomorrow, except tomorrow will be after a new quick change drive cage arrives
Some possibilities you could consider are (1) fewer partitions and (2) LVM.
LVM was beta at best the last time I tried it back in the fog of time when I had a little color left in my beard, and wound up losing the system. Is it trustable today?
Otherwise I'll probably just set it up as 2Gb for /boot, 500 Gb for /home (thats 4x the 127G its using now and the rest as /, which has actually worked pretty good for wheezy. Even that may be too complex, so /boot, 16G of /swap and the rest as / would also work. I had thought saveing /home separately might be a good idea and will likely do that now that I've surveyed the system with du -h. I'm about to take a 1T out, and put a 2T in for amanda at the same time, which would put a 1T available for /opt or even /home, but that drive has 75,000 spinning hours on it now. Its apparently found an immortality pill, it had 25 re-allocated sectors before I updated its firmware at about 1000 hours, still does! As vtapes for amanda, its stayed around 85% used for years, with 25 to 60 GB rewritten every night. Its also backing up my cnc machines too, 5 total.
Thinking out loud...
--Mike
To unsubscribe, e-mail: trinity-users-unsubscribe@lists.pearsoncomputing.net For additional commands, e-mail: trinity-users-help@lists.pearsoncomputing.net Read list messages on the web archive: http://trinity-users.pearsoncomputing.net/ Please remember not to top-post: http://trinity.pearsoncomputing.net/mailing_lists/#top-posting
LVM is better but it still makes recovery a huge pain vs using regular partitions. It's doable now but read up on it before you need it. Nowadays with larger drives very few people create multiple partitions unless you know you need them for a specific purpose. Create a small boot partition and dump everything else in root.
On Fri, Aug 24, 2018, 8:10 AM Gene Heskett gheskett@shentel.net wrote:
On Thursday 23 August 2018 23:24:37 Mike Bird wrote:
On Thu August 23 2018 19:52:22 Gene Heskett wrote:
And after about 3 more installs, I got it to work, but in my calculations for partition size, I forgot theres around 80 GB in /var, my web pages etc, which was going to be in the / partition, so 40Gb was soon filled up in copying it across. But lots of empty space left in home, so I fired up gparted to shrink the top of home and make room for a new /opt partition. But gparted refuses to do that resize/move. So it looks like another whole new install tomorrow, except tomorrow will be after a new quick change drive cage arrives
Some possibilities you could consider are (1) fewer partitions and (2) LVM.
LVM was beta at best the last time I tried it back in the fog of time when I had a little color left in my beard, and wound up losing the system. Is it trustable today?
Otherwise I'll probably just set it up as 2Gb for /boot, 500 Gb for /home (thats 4x the 127G its using now and the rest as /, which has actually worked pretty good for wheezy. Even that may be too complex, so /boot, 16G of /swap and the rest as / would also work. I had thought saveing /home separately might be a good idea and will likely do that now that I've surveyed the system with du -h. I'm about to take a 1T out, and put a 2T in for amanda at the same time, which would put a 1T available for /opt or even /home, but that drive has 75,000 spinning hours on it now. Its apparently found an immortality pill, it had 25 re-allocated sectors before I updated its firmware at about 1000 hours, still does! As vtapes for amanda, its stayed around 85% used for years, with 25 to 60 GB rewritten every night. Its also backing up my cnc machines too, 5 total.
Thinking out loud...
--Mike
To unsubscribe, e-mail: trinity-users-unsubscribe@lists.pearsoncomputing.net For additional commands, e-mail: trinity-users-help@lists.pearsoncomputing.net Read list messages on the web archive: http://trinity-users.pearsoncomputing.net/ Please remember not to top-post: http://trinity.pearsoncomputing.net/mailing_lists/#top-posting
-- Cheers, Gene Heskett -- "There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." -Ed Howdershelt (Author) Genes Web page http://geneslinuxbox.net:6309/gene
To unsubscribe, e-mail: trinity-users-unsubscribe@lists.pearsoncomputing.net For additional commands, e-mail: trinity-users-help@lists.pearsoncomputing.net Read list messages on the web archive: http://trinity-users.pearsoncomputing.net/ Please remember not to top-post: http://trinity.pearsoncomputing.net/mailing_lists/#top-posting
On Friday 24 August 2018 08:13:14 John Pisini wrote:
LVM is better but it still makes recovery a huge pain vs using regular partitions. It's doable now but read up on it before you need it. Nowadays with larger drives very few people create multiple partitions unless you know you need them for a specific purpose. Create a small boot partition and dump everything else in root.
On Fri, Aug 24, 2018, 8:10 AM Gene Heskett gheskett@shentel.net
wrote:
On Thursday 23 August 2018 23:24:37 Mike Bird wrote:
On Thu August 23 2018 19:52:22 Gene Heskett wrote:
And after about 3 more installs, I got it to work, but in my calculations for partition size, I forgot theres around 80 GB in /var, my web pages etc, which was going to be in the / partition, so 40Gb was soon filled up in copying it across. But lots of empty space left in home, so I fired up gparted to shrink the top of home and make room for a new /opt partition. But gparted refuses to do that resize/move. So it looks like another whole new install tomorrow, except tomorrow will be after a new quick change drive cage arrives
Some possibilities you could consider are (1) fewer partitions and (2) LVM.
LVM was beta at best the last time I tried it back in the fog of time when I had a little color left in my beard, and wound up losing the system. Is it trustable today?
The consensus is that LVM still isn't quite ready for prime time, and with big drives, no longer needed.
[...]
I'm about to take a 1T out, and put a 2T in for amanda at the same time, which would put a 1T available for /opt or even /home, but that drive has 75,000 spinning hours on it now. Its apparently found an immortality pill, it had 25 re-allocated sectors before I updated its firmware at about 1000 hours, still does! As vtapes for amanda, its stayed around 85% used for years, with 25 to 60 GB rewritten every night. Its also backing up my cnc machines too, 5 total.
Thinking out loud...
Thats done but with 2x the drive space now for amanda's vtapes, I have doubled the number of vtapes, and reconfigured it for that, and a starter session is running to shake it down & see if any bugs fall out.
This is of course via the wrapper scripts I wrote years ago because the normal usage's recovery loses a days work. When my version is done, I can do a bare metal recovery to the state the system is in when the dump of that disklist entry was done, so a much smaller window for data loss doing it my way. The config that made that backup run, and the indices of everything than run did, is now appended to the vtape its making, after its been made. So the recovery, if I need to do it yet today, will be effectively a state about an hour into the future a/o right now. I takes at least and hour to run, but if gzip is heavily used, will be even longer, 3 hours & change a few times. Amanda schedules differently from other backup proggies, fiddling with its schedule to try and make each run consume about the same amount of tape on each run, but remaining within the limits of how long it can go between level 0's. Since this will be effectively the first run, I doubt it will be able to do a level-0 on the number of 35GB vtapes I gave it as the maximum it can use per run. The email I get will be "interesting". But amanda is a smart girl, so a week later, she'll be doing alright.
Those scripts are available on my web pages if anyone else wants to play with them.
The other 2T drive has been reconfigured for the next stretch install, but I'll not try another install until amanda is done, middle of the afternoon earliest estimate.
Thanks John.
On Fri August 24 2018 09:09:14 Gene Heskett wrote:
The consensus is that LVM still isn't quite ready for prime time, and with big drives, no longer needed.
That is an opinion and also the consensus as of 15-20 years ago.
It is not the current consensus.
We have used LVM since 2004 or before. It would take too long for me to determine the exact start date and the total number of systems upon which we have deployed LVM. I can however quote you from memory the number of problems we have experienced: zero.
For simple systems it is generally easiest to use only a boot and an "everything else" partition with neither LVM nor complex partitioning.
For more complex systems LVM is a valuable and robust tool and much more flexible than partitioning. Within a single complex system we may use multiple volume groups with different PE sizes, different RAID levels, different block/inode ratios, different reserved block percentages, different mount attributes (e.g. noexec), and different user quotas. We currently use ext3 exclusively but others may also use different filesystem types in different logical volumes as appropriate.
FWIW we have many times found LVM helpful when migrating from failing hard drives to new drives - just add the new physical volumes, remove the old physical volumes, and everything is migrated by magic.
--Mike
On Friday 24 of August 2018 19:23:09 Mike Bird wrote:
On Fri August 24 2018 09:09:14 Gene Heskett wrote:
The consensus is that LVM still isn't quite ready for prime time, and with big drives, no longer needed.
That is an opinion and also the consensus as of 15-20 years ago.
It is not the current consensus.
We have used LVM since 2004 or before. It would take too long for me to determine the exact start date and the total number of systems upon which we have deployed LVM. I can however quote you from memory the number of problems we have experienced: zero.
For simple systems it is generally easiest to use only a boot and an "everything else" partition with neither LVM nor complex partitioning.
For more complex systems LVM is a valuable and robust tool and much more flexible than partitioning. Within a single complex system we may use multiple volume groups with different PE sizes, different RAID levels, different block/inode ratios, different reserved block percentages, different mount attributes (e.g. noexec), and different user quotas. We currently use ext3 exclusively but others may also use different filesystem types in different logical volumes as appropriate.
FWIW we have many times found LVM helpful when migrating from failing hard drives to new drives - just add the new physical volumes, remove the old physical volumes, and everything is migrated by magic.
--Mike
Yes, just under that I would sign!
For me too, LVM is an indispensable tool. In addition to managing filesystems, in my case, LVM is also an indispensable tool for virtual machine drives (of course I use Debian with Linux KVM as hypervisor).
In the years when I use it, I remember only one problem. But that's a long past.
Cheers
On Thursday 23 August 2018 07:21:47 Stefan Krusche wrote:
wget -O keyfile 'http://keyserver.quickbuild.io:11371/pks/lookup?op=get&search=0xF5CFC 95C
The above, with the added quote, executed here of this wheezy install;
gene@coyote:/$ wget -O keyfile 'http://keyserver.quickbuild.io:11371/pks/lookup?op=get&search=0xF5CFC95C' Cannot specify both -k and -O if multiple URLs are given, or in combination with -p or -r. See the manual for details.
Usage: wget [OPTION]... [URL]...
And of course not even the empty keyfile is generated.
alias returns: alias gen-password='head -c 21 /dev/urandom | base64 | head -c 26 && echo' alias linuxcnc='xrdb -all -query|sed -e '''s#[A-Z_]*BACKGROUND# gray90#''' -e '''s#[A-Z_]*FOREGROUND# Black#''' -e '''s#[A-Z_]*HIGHLIGHT# White#''' -e '''s#[A-Z_]*LOWLIGHT# Black#'''|xrdb -merge; linuxcnc' alias ls='ls --color=auto' alias nano='nano --smooth '
Now, just for S&G I ssh'd into an intel box on my locate network, AND IT WORKED!!!!!!
Nothing there that should effect wget that I can see. And the re-install of wget isn't enabled by synaptic. I just ran that line on one of my milling machines, and now have a keyfile, it worked so the wget on this machine is fubar. So once I've done another install, I will see if I can get it copied back and installed here. However, I should point out that this is a 3rd copy of wget thats fubar since its the stretch version of apt-key that fails with the 'connection refused' error.
And its not wget, the sha1sums are the same on both machines. one works, the other doesn't. But both are wheezy
Is it 17:00 yet?
On 08/23/2018 07:42 AM, Gene Heskett wrote:
gene@coyote:/$ wget -O keyfile 'http://keyserver.quickbuild.io:11371/pks/lookup?op=get&search=0xF5CFC95C' Cannot specify both -k and -O if multiple URLs are given, or in combination with -p or -r. See the manual for details.
I think the word 'keyfile' shouldn't be there. As far as I know it's not a valid wget command, so wget is interpreting it as the option -k (--convert-links).
On Thu August 23 2018 07:56:45 Dan Youngquist wrote:
On 08/23/2018 07:42 AM, Gene Heskett wrote:
gene@coyote:/$ wget -O keyfile 'http://keyserver.quickbuild.io:11371/pks/lookup?op=get&search=0xF5CFC95C ' Cannot specify both -k and -O if multiple URLs are given, or in combination with -p or -r. See the manual for details.
I think the word 'keyfile' shouldn't be there. As far as I know it's not a valid wget command, so wget is interpreting it as the option -k (--convert-links).
"-O keyfile" tells wget to put the key (plus some HTML) into the file "keyfile".
--Mike
On Thursday 23 August 2018 10:56:45 Dan Youngquist wrote:
On 08/23/2018 07:42 AM, Gene Heskett wrote:
gene@coyote:/$ wget -O keyfile 'http://keyserver.quickbuild.io:11371/pks/lookup?op=get&search=0xF5C FC95C' Cannot specify both -k and -O if multiple URLs are given, or in combination with -p or -r. See the manual for details.
I think the word 'keyfile' shouldn't be there. As far as I know it's not a valid wget command, so wget is interpreting it as the option -k (--convert-links).
re-read the manpage.
"-O keyfile"
is the filename to store the collected key into. I wondered about that myself so I looked it up. :)
Gene Heskett composed on 2018-08-21 14:47 (UTC-0400):
No, wheezy is still in the main repos, it have not yet moved to archive.
I took the ".us" out of the hostnames, now my fail list is 2x longer, no hostname associated failures. If they still exist, I need the full url's. please?
Right out of sources.list:
http://ftp.us.debian.org/debian/dists/wheezy/
URLs copied from Firefox and pasted.
On Tuesday 21 August 2018 15:13:52 Felix Miata wrote:
Gene Heskett composed on 2018-08-21 14:47 (UTC-0400):
No, wheezy is still in the main repos, it have not yet moved to archive.
I took the ".us" out of the hostnames, now my fail list is 2x longer, no hostname associated failures. If they still exist, I need the full url's. please?
Right out of sources.list:
extend it with "dists/wheezy" and it works, but nothing there I don't already have.
Thanks Felix.
http://ftp.us.debian.org/debian/dists/wheezy/
URLs copied from Firefox and pasted.
On Tue, 21 Aug 2018 14:47:35 -0400 Gene Heskett gheskett@shentel.net wrote:
On Tuesday 21 August 2018 13:43:18 Nick Koretsky wrote:
On Tue, 21 Aug 2018 13:18:38 -0400
Gene Heskett gheskett@shentel.net wrote:
On Tue, 21 Aug 2018 10:10:56 -0400
Gene Heskett gheskett@shentel.net wrote:
From the plethora of errors I'm getting in a synaptic refresh, the wheezy repos seem to have been shut down. Support ended June 30 IIRC. Even backports is 404. The rest are address not found.
Hm? http://ftp.debian.org/ still have wheezy.
but no key files, something aptitude doesn't tolerate, and synaptic complains about at length. or just plain gone. If synaptic will copy/paste: Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/Release.gpg Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/main/binary- i386/Packages Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/contrib/bina ry-i386/Packages Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/non-free/bin ary-i386/Packages Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/contrib/i18n /Translation-en_US Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/contrib/i18n /Translation-en Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/main/i18n/Tr anslation-en_US Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/main/i18n/Tr anslation-en Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/non-free/i18 n/Translation-en_US Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://http.us.debian.net/debian/dists/wheezy-backports/non-free/i18 n/Translation-en Something wicked happened resolving 'http.us.debian.net:http' (-5 - No address associated with hostname) Failed to fetch http://mozilla.debian.net/dists/wheezy-backports/firefox-release/bin ary-i386/Packages 404 Not Found [IP: 149.20.4.15 80] Some index files failed to download. They have been ignored, or old ones used instead.
The us.debian.net mirror is down apparently. Use a different one. I have a wheezy machine and just ran a apt-get update without problem.
The debian repos are never gone, the older one just get moved to http://archive.debian.org
So I should edit the repos to be archive.debian.org? Is there just one, no mirrors?
No, wheezy is still in the main repos, it have not yet moved to archive.
I took the ".us" out of the hostnames, now my fail list is 2x longer, no hostname associated failures. If they still exist, I need the full url's. please?
deb http://ftp.debian.org/debian wheezy main contrib non-free deb-src http://ftp.debian.org/debian wheezy main contrib non-free
deb http://ftp.debian.org/debian wheezy-updates main contrib non-free deb-src http://ftp.debian.org/debian wheezy-updates main contrib non-free
deb http://security.debian.org/ wheezy/updates main contrib non-free deb-src http://security.debian.org/ wheezy/updates main contrib non-free
deb http://ftp.debian.org/debian wheezy-backports main contrib non-free deb-src http://ftp.debian.org/debian wheezy-backports main contrib non-free
On 08/21/2018 11:47 AM, Gene Heskett wrote:
On Tuesday 21 August 2018 13:43:18 Nick Koretsky wrote:
On Tue, 21 Aug 2018 13:18:38 -0400
Gene Heskett gheskett@shentel.net wrote:
On Tue, 21 Aug 2018 10:10:56 -0400
Gene Heskett gheskett@shentel.net wrote:
From the plethora of errors I'm getting in a synaptic refresh, the wheezy repos seem to have been shut down. Support ended June 30 IIRC. Even backports is 404. The rest are address not found.
Hm? http://ftp.debian.org/ still have wheezy.
I got recent updates and a new kernel using these sources...
deb http://ftp.us.debian.org/debian/ wheezy main contrib non-free deb http://security.debian.org/ wheezy/updates main contrib non-free deb http://ftp.us.debian.org/debian/ wheezy-updates main contrib non-free deb http://ftp.debian.org/debian/ wheezy-backports main contrib non-free
On Wednesday 22 of August 2018 05:43:39 Jimmy Johnson wrote:
On 08/21/2018 11:47 AM, Gene Heskett wrote:
On Tuesday 21 August 2018 13:43:18 Nick Koretsky wrote:
On Tue, 21 Aug 2018 13:18:38 -0400
Gene Heskett gheskett@shentel.net wrote:
On Tue, 21 Aug 2018 10:10:56 -0400
Gene Heskett gheskett@shentel.net wrote:
From the plethora of errors I'm getting in a synaptic refresh, the wheezy repos seem to have been shut down. Support ended June 30 IIRC. Even backports is 404. The rest are address not found.
Hm? http://ftp.debian.org/ still have wheezy.
I got recent updates and a new kernel using these sources...
deb http://ftp.us.debian.org/debian/ wheezy main contrib non-free deb http://security.debian.org/ wheezy/updates main contrib non-free deb http://ftp.us.debian.org/debian/ wheezy-updates main contrib non-free deb http://ftp.debian.org/debian/ wheezy-backports main contrib non-free
I recommend using Debian redirector instead of a particular server - in the rows above, replace ftp.us.debian.org with deb.debian.org
Cheers
On Wednesday 22 August 2018 02:59:30 Slávek Banko wrote:
On Wednesday 22 of August 2018 05:43:39 Jimmy Johnson wrote:
On 08/21/2018 11:47 AM, Gene Heskett wrote:
On Tuesday 21 August 2018 13:43:18 Nick Koretsky wrote:
On Tue, 21 Aug 2018 13:18:38 -0400
Gene Heskett gheskett@shentel.net wrote:
On Tue, 21 Aug 2018 10:10:56 -0400
Gene Heskett gheskett@shentel.net wrote: > From the plethora of errors I'm getting in a synaptic > refresh, the wheezy repos seem to have been shut down. Support > ended June 30 IIRC. Even backports is 404. The rest are > address not found.
Hm? http://ftp.debian.org/ still have wheezy.
I got recent updates and a new kernel using these sources...
deb http://ftp.us.debian.org/debian/ wheezy main contrib non-free deb http://security.debian.org/ wheezy/updates main contrib non-free deb http://ftp.us.debian.org/debian/ wheezy-updates main contrib non-free deb http://ftp.debian.org/debian/ wheezy-backports main contrib non-free
I recommend using Debian redirector instead of a particular server - in the rows above, replace ftp.us.debian.org with deb.debian.org
Cheers
Failed to fetch http://deb.debian.org/debian/dists/dists/wheezy/main/binary-i386/Packages 404 Not Found Failed to fetch http://deb.debian.org/debian/dists/dists/wheezy/contrib/binary-i386/Packages 404 Not Found Failed to fetch http://deb.debian.org/debian/dists/dists/wheezy/non-free/binary-i386/Package... 404 Not Found Failed to fetch http://deb.debian.org/debian/dists/dists/wheezy-updates/main/binary-i386/Pac... 404 Not Found Failed to fetch http://deb.debian.org/debian/dists/dists/wheezy-updates/contrib/binary-i386/... 404 Not Found Failed to fetch http://deb.debian.org/debian/dists/dists/wheezy-updates/non-free/binary-i386... 404 Not Found
Got rid of the extra "/dists/ that someone said I needed, and it all works except anything not installed is now "new in repository". A sea of blue I haven't seen before.
Am Dienstag, 21. August 2018 schrieb Slávek Banko:
Dne út 21. srpna 2018 Nick Koretsky napsal(a):
On Tue, 21 Aug 2018 09:08:55 -0400
Gene Heskett gheskett@shentel.net wrote:
On Tuesday 21 August 2018 08:21:12 Nick Koretsky wrote:
On Tue, 21 Aug 2018 12:55:28 +0200
"Dr. Nikolaus Klepp" office@klepp.biz wrote:
Am Dienstag, 21. August 2018 schrieb Nick Koretsky:
On Tue, 21 Aug 2018 09:20:43 +0200
"Dr. Nikolaus Klepp" office@klepp.biz wrote: > > Its tears so far Nik. There are so many differences just in > > the ext4's used that one of them should be renamed, they > > will NOT cross mount, ext4 disk to ext4 mount. > > Hi Gene! > > When you cannot mount the ext4 partions from one another, > then there is something very wrong. ext4 can be mounted as > ext2 and that should alway work - at least, if the drives and > filesystems are ok.
No, you are wrong. There were options added to ext4 which made it incompatible with older kernels. And a few years ago they made this options default. Debian wheezy kernel (3.2) would not mount ext4 created in debian stretch.
Hi!
Now that's a gotcha I did not know. How did it come that this was thought of beeing a good idea?
Nik
Yep. Exactly my thought when i spent 3-hours with a server refusing to boot after migration to a new hdd (i used a stretch flash to copy). Who the fuck toughs it was a good idea?!! Why not call it ext5 or ext4a or whatever?!!
+100
Now the question of the day is how the hell do we fix it?
If you need a shared partition between wheezy and stretch create it in wheezy.
I recommend upgrading to wheezy-backports kernel.
Cheers
Maybe I missed the point somewhere, but I was under the impression that Gene wants to run linuxcnc on a production machine, and that needs a realtime kernel. To be more prcise, linuxcnc 2.7 needs a rtai patched kernel if you want decent latentcy, not the rtpreempt kernels from debian. I have now two mills running on devuan ascii, both with the old kernels from linuxcnc-wheezy 'cause of latency.
Nik
On Tuesday 21 August 2018 16:55:31 Dr. Nikolaus Klepp wrote:
Am Dienstag, 21. August 2018 schrieb Slávek Banko:
Dne út 21. srpna 2018 Nick Koretsky napsal(a):
On Tue, 21 Aug 2018 09:08:55 -0400
Gene Heskett gheskett@shentel.net wrote:
On Tuesday 21 August 2018 08:21:12 Nick Koretsky wrote:
On Tue, 21 Aug 2018 12:55:28 +0200
"Dr. Nikolaus Klepp" office@klepp.biz wrote:
Am Dienstag, 21. August 2018 schrieb Nick Koretsky: > On Tue, 21 Aug 2018 09:20:43 +0200 > > "Dr. Nikolaus Klepp" office@klepp.biz wrote: > > > Its tears so far Nik. There are so many differences > > > just in the ext4's used that one of them should be > > > renamed, they will NOT cross mount, ext4 disk to ext4 > > > mount. > > > > Hi Gene! > > > > When you cannot mount the ext4 partions from one > > another, then there is something very wrong. ext4 can be > > mounted as ext2 and that should alway work - at least, > > if the drives and filesystems are ok. > > No, you are wrong. There were options added to ext4 which > made it incompatible with older kernels. And a few years > ago they made this options default. Debian wheezy kernel > (3.2) would not mount ext4 created in debian stretch.
Hi!
Now that's a gotcha I did not know. How did it come that this was thought of beeing a good idea?
Nik
Yep. Exactly my thought when i spent 3-hours with a server refusing to boot after migration to a new hdd (i used a stretch flash to copy). Who the fuck toughs it was a good idea?!! Why not call it ext5 or ext4a or whatever?!!
+100
Now the question of the day is how the hell do we fix it?
If you need a shared partition between wheezy and stretch create it in wheezy.
I recommend upgrading to wheezy-backports kernel.
Cheers
Maybe I missed the point somewhere, but I was under the impression that Gene wants to run linuxcnc on a production machine, and that needs a realtime kernel. To be more prcise, linuxcnc 2.7 needs a rtai patched kernel if you want decent latentcy, not the rtpreempt kernels from debian. I have now two mills running on devuan ascii, both with the old kernels from linuxcnc-wheezy 'cause of latency.
Nik
Glad I have a rowing partner in this leaky rowboat, Nik.
On Tuesday 21 August 2018 16:55:31 Dr. Nikolaus Klepp wrote:
Am Dienstag, 21. August 2018 schrieb Slávek Banko:
Dne út 21. srpna 2018 Nick Koretsky napsal(a):
On Tue, 21 Aug 2018 09:08:55 -0400
Gene Heskett gheskett@shentel.net wrote:
On Tuesday 21 August 2018 08:21:12 Nick Koretsky wrote:
On Tue, 21 Aug 2018 12:55:28 +0200
"Dr. Nikolaus Klepp" office@klepp.biz wrote:
Am Dienstag, 21. August 2018 schrieb Nick Koretsky: > On Tue, 21 Aug 2018 09:20:43 +0200 > > "Dr. Nikolaus Klepp" office@klepp.biz wrote: > > > Its tears so far Nik. There are so many differences > > > just in the ext4's used that one of them should be > > > renamed, they will NOT cross mount, ext4 disk to ext4 > > > mount. > > > > Hi Gene! > > > > When you cannot mount the ext4 partions from one > > another, then there is something very wrong. ext4 can be > > mounted as ext2 and that should alway work - at least, > > if the drives and filesystems are ok. > > No, you are wrong. There were options added to ext4 which > made it incompatible with older kernels. And a few years > ago they made this options default. Debian wheezy kernel > (3.2) would not mount ext4 created in debian stretch.
Hi!
Now that's a gotcha I did not know. How did it come that this was thought of beeing a good idea?
Nik
Yep. Exactly my thought when i spent 3-hours with a server refusing to boot after migration to a new hdd (i used a stretch flash to copy). Who the fuck toughs it was a good idea?!! Why not call it ext5 or ext4a or whatever?!!
+100
Now the question of the day is how the hell do we fix it?
If you need a shared partition between wheezy and stretch create it in wheezy.
I recommend upgrading to wheezy-backports kernel.
Cheers
Maybe I missed the point somewhere, but I was under the impression that Gene wants to run linuxcnc on a production machine, and that needs a realtime kernel. To be more prcise, linuxcnc 2.7 needs a rtai patched kernel if you want decent latentcy, not the rtpreempt kernels from debian. I have now two mills running on devuan ascii, both with the old kernels from linuxcnc-wheezy 'cause of latency.
Nik
My last 2 or 3 emails seem to have fallen off the edge of the planet. Anyway I have now swapped the drives around and used the bios boot menu to select a boot from what is now sdB, and its working. Next is to use this copy of gparted to reformat /dev/sdA and then see if a successful install can be done. I'd say buy me a beer but that might royally screw things up, so wish me luck (good that is) ;-) This will be about the 6th full install today. But before I reboot to do that, I'd better fix us some din-din. Maybe someone will have a better idea by then?
Am Mittwoch, 22. August 2018 schrieb Gene Heskett:
[...] My last 2 or 3 emails seem to have fallen off the edge of the planet. Anyway I have now swapped the drives around and used the bios boot menu to select a boot from what is now sdB, and its working. Next is to use this copy of gparted to reformat /dev/sdA and then see if a successful install can be done. I'd say buy me a beer but that might royally screw things up, so wish me luck (good that is) ;-) This will be about the 6th full install today. But before I reboot to do that, I'd better fix us some din-din. Maybe someone will have a better idea by then?
Hi Gene!
My recipe to Devuan ascii + LinuxCNC is this (take the commandlines with a salt of grain, I'm not the machine right now): - Install linuxcnc from the linuxcnc ISO - "aptitude update && aptitude dist-upgrade" - "apt-mark hold linux-image-3...." - "apt-mark hold rtai-...." - "apt-mark hold linuxcnc" - add Devuan repsitory for jessie, upgrade to jessie. - "aptitude purge $(dpkg -l|grep ^rc|tr : ' '|awk '{print $2}') - change from jessie to ascii, upgrade to asci - install "apt-show-versions" - aptitude purge $(apt-show-versions|grep --invert-match uptodate|cut : ' '| awk '{print $1}' - when this is done, I intall TDE and purge all gnome stuff, and install NCAM After this operation "axis" does not work, but "gmoccapy". "ncam" works, "camview" does not, but "mplayer" does.
and get a bit more bottels of beer ..
Nik
On Wednesday 22 August 2018 04:32:31 Dr. Nikolaus Klepp wrote:
Am Mittwoch, 22. August 2018 schrieb Gene Heskett:
[...] My last 2 or 3 emails seem to have fallen off the edge of the planet. Anyway I have now swapped the drives around and used the bios boot menu to select a boot from what is now sdB, and its working. Next is to use this copy of gparted to reformat /dev/sdA and then see if a successful install can be done. I'd say buy me a beer but that might royally screw things up, so wish me luck (good that is) ;-) This will be about the 6th full install today. But before I reboot to do that, I'd better fix us some din-din. Maybe someone will have a better idea by then?
Hi Gene!
My recipe to Devuan ascii + LinuxCNC is this (take the commandlines with a salt of grain, I'm not the machine right now): - Install linuxcnc from the linuxcnc ISO
- "aptitude update && aptitude dist-upgrade"
- "apt-mark hold linux-image-3...."
- "apt-mark hold rtai-...."
- "apt-mark hold linuxcnc"
- add Devuan repsitory for jessie, upgrade to jessie.
- "aptitude purge $(dpkg -l|grep ^rc|tr : ' '|awk '{print $2}')
- change from jessie to ascii, upgrade to asci
- install "apt-show-versions"
- aptitude purge $(apt-show-versions|grep --invert-match uptodate|cut
: ' '| awk '{print $1}' - when this is done, I intall TDE and purge all gnome stuff, and install NCAM After this operation "axis" does not work, but "gmoccapy". "ncam" works, "camview" does not, but "mplayer" does.
I'm stuck on axis, and all of the extra pyvcp/xml stuff I've done depends on axis. I took ncam out of synaptic, didn't use it, I write my own gcode 99% of the time anyway, and got tired of the updates being moved about monthly. So you have to login to the forum, find Fern's messages and edit the repo line before a new one can be had. And the last one didn't want to install. I didn't investigate why in much depth. I think Fern is running a newer distro now so dependencies on newer stuff kills it.
And camview has never recovered from being asked to run on wheezy. It will run, until the too small axis window is expanded but if its window goes above the cameras native format size width, it goes away and the only way to restore it is to reboot, a logout/back in is not sufficient. But cheese still works. Go figure but I've now duplicated it with 3 different cameras. To me, since thats written in python2, its a python bug. No clue if porting it to python-3.7 might help.
and get a bit more bottels of beer ..
Local stores are out of a beer a diabetic can drink, Miller64, or even select55, but that stuffs been thru the horse too many times, no taste left for me, and have been for about 2 weeks now. I think its the distributer slacking off. And I'd have to drive 100 miles to get out of his area. I usually drink one for dinner, ostensibly to wash down a gram of metformin I take 2 a day of for my sugar.
Nik
Thanks Nik.
On 08/21/2018 06:36 AM, Nick Koretsky wrote:
On Tue, 21 Aug 2018 09:08:55 -0400 Gene Heskett wrote:
Now the question of the day is how the hell do we fix it?
If you need a shared partition between wheezy and stretch create it in wheezy.
Better yet, don't use ext4. Years ago I considered switching to ext4, so looked at it compared to ext3, and concluded that ext4 has a slightly greater chance of data loss. So I stayed with ext3. I'm now just beginning to gradually switch over to btrfs because of its defrag capability.
Gene Heskett composed on 2018-08-21 09:08 (UTC-0400):
Now that's a gotcha I did not know. How did it come that this was thought of beeing a good idea?
Yep. Exactly my thought when i spent 3-hours with a server refusing to boot after migration to a new hdd (i used a stretch flash to copy). Who the fuck toughs it was a good idea?!! Why not call it ext5 or ext4a or whatever?!!
+100
Now the question of the day is how the hell do we fix it?
As I implied 24 hours ago in this thread, it doesn't need fixing. Jessie came in between Wheezy and Stretch, a long time (two years?). There are multiple workarounds for those who skipped over Jessie:
1-Format the installation target using Wheezy or Jessie or old Knoppix or old anything in advance of starting the installer. Formatting is an /optional/ part of the installation process (as is partitioning).
2-As Dan wrote, use EXT3 instead of EXT4 (EXT3 can be converted to EXT4 later if desired. AFAIK, the incompatible options will not automatically be added on conversion.)
3-Format with the Stretch installer, but specify to omit the incompatible EXT4 options. I don't which it/they are, but my guess is the only one that is needed is "-O '^64bit'". All my Wheezys were eliminated around the time its support termination was announced, so no practical way here to test.
4-Use XFS instead of EXT4.
5-Use a newer and/or custom kernel in Wheezy. http://ftp5.gwdg.de/pub/linux/debian/debian/pool/main/l/
On Tuesday 21 August 2018 10:56:56 Felix Miata wrote:
Gene Heskett composed on 2018-08-21 09:08 (UTC-0400):
Now that's a gotcha I did not know. How did it come that this was thought of beeing a good idea?
Yep. Exactly my thought when i spent 3-hours with a server refusing to boot after migration to a new hdd (i used a stretch flash to copy). Who the fuck toughs it was a good idea?!! Why not call it ext5 or ext4a or whatever?!!
+100
Now the question of the day is how the hell do we fix it?
As I implied 24 hours ago in this thread, it doesn't need fixing. Jessie came in between Wheezy and Stretch, a long time (two years?). There are multiple workarounds for those who skipped over Jessie:
1-Format the installation target using Wheezy or Jessie or old Knoppix or old anything in advance of starting the installer. Formatting is an /optional/ part of the installation process (as is partitioning).
I did try to do that, it showed me the partitions I had setup in gparted, but would not skip the selection and format steps.
2-As Dan wrote, use EXT3 instead of EXT4 (EXT3 can be converted to EXT4 later if desired. AFAIK, the incompatible options will not automatically be added on conversion.)
3-Format with the Stretch installer, but specify to omit the incompatible EXT4 options. I don't which it/they are, but my guess is the only one that is needed is "-O '^64bit'". All my Wheezys were eliminated around the time its support termination was announced, so no practical way here to test.
4-Use XFS instead of EXT4.
Thats been contemplated, but not being able to skip the partition and formatting stage, I don't know how.
If there is a way, plz advise.
But this morning, the first install wouldn't boot at, as if there was no grub on the drive, but I watched it install it. All I could see was no drive activity, and a blinking underline cursor in row/column 1. Several resets and a choose /sdb to boot from each time. So I started another install, non-graphical, which sorta worked below.
2nd try today I only told it a separate /home, and that seemed to work, but despite not choosing that in the initial menu, the speech-dispatcher drove me crazy. No volume control installed, and the keyboard echo to the terminal became so slow it screwed up my typing when I had become root and issued a killall to shut it up, as if it was waiting for the speech-dispatcher to finish. A gracefull stop as an option to the script in /etc/init.d was ignored, as root or as me.
Hell of a way to run a train.
I added the wiki's entries in sources.list.d/trinity.list, but then the keyserver refused my connection about 15 times so now I'm back on wheezy since I can't replace the xfce default.
So whats with keyserver.quickbuild.io this morning, I can ping it from stretch just fine. And the missus wants the papers before the storm arrives, so I'll do that next. She's a good girl but a Crossword addict.
5-Use a newer and/or custom kernel in Wheezy. http://ftp5.gwdg.de/pub/linux/debian/debian/pool/main/l/
I recently had to reinstall (I am currently staying on Jessie it does pretty much everything I need and the bugs that exist I know how to work around) and I had a similar issue. Grub would install but wouldn't boot the machine. I could boot on my machine f11 to choose which device and I could see the right loader and actually use it to boot the machine but I could not get the machine to move that to the primary spot. in my case as I had full backups and it was a new install with no real customization I just completely flattened the disk and started over.
On Tue, Aug 21, 2018 at 12:06 PM, Gene Heskett gheskett@shentel.net wrote:
On Tuesday 21 August 2018 10:56:56 Felix Miata wrote:
Gene Heskett composed on 2018-08-21 09:08 (UTC-0400):
Now that's a gotcha I did not know. How did it come that this was thought of beeing a good idea?
Yep. Exactly my thought when i spent 3-hours with a server refusing to boot after migration to a new hdd (i used a stretch flash to copy). Who the fuck toughs it was a good idea?!! Why not call it ext5 or ext4a or whatever?!!
+100
Now the question of the day is how the hell do we fix it?
As I implied 24 hours ago in this thread, it doesn't need fixing. Jessie came in between Wheezy and Stretch, a long time (two years?). There are multiple workarounds for those who skipped over Jessie:
1-Format the installation target using Wheezy or Jessie or old Knoppix or old anything in advance of starting the installer. Formatting is an /optional/ part of the installation process (as is partitioning).
I did try to do that, it showed me the partitions I had setup in gparted, but would not skip the selection and format steps.
2-As Dan wrote, use EXT3 instead of EXT4 (EXT3 can be converted to EXT4 later if desired. AFAIK, the incompatible options will not automatically be added on conversion.)
3-Format with the Stretch installer, but specify to omit the incompatible EXT4 options. I don't which it/they are, but my guess is the only one that is needed is "-O '^64bit'". All my Wheezys were eliminated around the time its support termination was announced, so no practical way here to test.
4-Use XFS instead of EXT4.
Thats been contemplated, but not being able to skip the partition and formatting stage, I don't know how.
If there is a way, plz advise.
But this morning, the first install wouldn't boot at, as if there was no grub on the drive, but I watched it install it. All I could see was no drive activity, and a blinking underline cursor in row/column 1. Several resets and a choose /sdb to boot from each time. So I started another install, non-graphical, which sorta worked below.
2nd try today I only told it a separate /home, and that seemed to work, but despite not choosing that in the initial menu, the speech-dispatcher drove me crazy. No volume control installed, and the keyboard echo to the terminal became so slow it screwed up my typing when I had become root and issued a killall to shut it up, as if it was waiting for the speech-dispatcher to finish. A gracefull stop as an option to the script in /etc/init.d was ignored, as root or as me.
Hell of a way to run a train.
I added the wiki's entries in sources.list.d/trinity.list, but then the keyserver refused my connection about 15 times so now I'm back on wheezy since I can't replace the xfce default.
So whats with keyserver.quickbuild.io this morning, I can ping it from stretch just fine. And the missus wants the papers before the storm arrives, so I'll do that next. She's a good girl but a Crossword addict.
5-Use a newer and/or custom kernel in Wheezy. http://ftp5.gwdg.de/pub/linux/debian/debian/pool/main/l/
-- Cheers, Gene Heskett -- "There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." -Ed Howdershelt (Author) Genes Web page http://geneslinuxbox.net:6309/gene
To unsubscribe, e-mail: trinity-users-unsubscribe@ lists.pearsoncomputing.net For additional commands, e-mail: trinity-users-help@lists. pearsoncomputing.net Read list messages on the web archive: http://trinity-users. pearsoncomputing.net/ Please remember not to top-post: http://trinity. pearsoncomputing.net/mailing_lists/#top-posting
Gene Heskett composed on 2018-08-21 12:06 (UTC-0400):
On Tuesday 21 August 2018 10:56:56 Felix Miata wrote:
As I implied 24 hours ago in this thread, it doesn't need fixing. Jessie came in between Wheezy and Stretch, a long time (two years?). There are multiple workarounds for those who skipped over Jessie:
1-Format the installation target using Wheezy or Jessie or old Knoppix or old anything in advance of starting the installer. Formatting is an /optional/ part of the installation process (as is partitioning).
I did try to do that, it showed me the partitions I had setup in gparted, but would not skip the selection and format steps.
I think you must be misunderstanding what these steps are. They must be used to specify which existing partitions to use, where to mount them, mount options, etc. Once these have been done with supported formatted empty partitions, the installer will default to "do not format". Even if they managed to return "format partition", you can change the selection to "do not format".
On Tuesday 21 August 2018 12:49:53 Felix Miata wrote:
Gene Heskett composed on 2018-08-21 12:06 (UTC-0400):
On Tuesday 21 August 2018 10:56:56 Felix Miata wrote:
As I implied 24 hours ago in this thread, it doesn't need fixing. Jessie came in between Wheezy and Stretch, a long time (two years?). There are multiple workarounds for those who skipped over Jessie:
1-Format the installation target using Wheezy or Jessie or old Knoppix or old anything in advance of starting the installer. Formatting is an /optional/ part of the installation process (as is partitioning).
I did try to do that, it showed me the partitions I had setup in gparted, but would not skip the selection and format steps.
I think you must be misunderstanding what these steps are. They must be used to specify which existing partitions to use, where to mount them, mount options, etc. Once these have been done with supported formatted empty partitions, the installer will default to "do not format". Even if they managed to return "format partition", you can change the selection to "do not format".
printed, I'll try it again. Thanks Felix.
On Tuesday 21 August 2018 10:56:56 Felix Miata wrote:
Gene Heskett composed on 2018-08-21 09:08 (UTC-0400):
Now that's a gotcha I did not know. How did it come that this was thought of beeing a good idea?
Yep. Exactly my thought when i spent 3-hours with a server refusing to boot after migration to a new hdd (i used a stretch flash to copy). Who the fuck toughs it was a good idea?!! Why not call it ext5 or ext4a or whatever?!!
+100
Now the question of the day is how the hell do we fix it?
As I implied 24 hours ago in this thread, it doesn't need fixing. Jessie came in between Wheezy and Stretch, a long time (two years?). There are multiple workarounds for those who skipped over Jessie:
With all due respect, and thats quite a bit for you Felix, there is but one Jessie install here that runs Linuxcnc, and its on a raspi so we're downloading a sd image and the install is writing the card and plugging it in. I tried the installer on x86 machines and failed quite a few times on several different machines. The LCNC folks never did issue a dvd image of jessie that ran LinuxCNC. Now they have an experimental stretch amd64 image that I'm playing with now, but IMO that installer is broken too, no way to shut the speech-dispatcher up short of turning off the audio amp. Had there ever been a working amd64 or even a 32 bit x86 jessie dvd image I would probably be running it on 6 machines here. That one jessie I have, running most of a ton of a 70 yo old Sheldon lathe from an rpi-3b, is dead stable from local power outage to local power outage. Too bad it never got a working installer for amd64.
1-Format the installation target using Wheezy or Jessie or old Knoppix or old anything in advance of starting the installer. Formatting is an /optional/ part of the installation process (as is partitioning).
BTDT, the installer can't bypass it anyway. It it can, how?
2-As Dan wrote, use EXT3 instead of EXT4 (EXT3 can be converted to EXT4 later if desired. AFAIK, the incompatible options will not automatically be added on conversion.)
3-Format with the Stretch installer, but specify to omit the incompatible EXT4 options. I don't which it/they are, but my guess is the only one that is needed is "-O '^64bit'". All my Wheezys were eliminated around the time its support termination was announced, so no practical way here to test.
So theres no lifeboat left in the wild. :(
4-Use XFS instead of EXT4.
5-Use a newer and/or custom kernel in Wheezy. http://ftp5.gwdg.de/pub/linux/debian/debian/pool/main/l/
I downloaded one of the amd64-4.18 kernels, but since this system is a 32 bit system, it likely won't install. One of the reasons I wanted to update was to get into 64 bit, but LCNC up until about a year ago, was a 32 bit app only because the latency went to hell with 64 bits, the context switch frame to service an interrupt was that much bigger.
But now nearly everyone including me has exported the realtime stuff to fpga cards which in some simpler cases become sub $100 expenditures instead of trying to pour all that data back and forth over an epp parport, latency is somewhat less important than formerly. Now its a balancing act between machine max speed and adequate control, usually in 1 millisecond windows.
Thanks Felix.
On Tuesday 21 August 2018 05:12:03 Nick Koretsky wrote:
On Tue, 21 Aug 2018 09:20:43 +0200
"Dr. Nikolaus Klepp" office@klepp.biz wrote:
Its tears so far Nik. There are so many differences just in the ext4's used that one of them should be renamed, they will NOT cross mount, ext4 disk to ext4 mount.
Hi Gene!
When you cannot mount the ext4 partions from one another, then there is something very wrong. ext4 can be mounted as ext2 and that should alway work - at least, if the drives and filesystems are ok.
No, you are wrong. There were options added to ext4 which made it incompatible with older kernels. And a few years ago they made this options default. Debian wheezy kernel (3.2) would not mount ext4 created in debian stretch.
But will it work in reverse by making the /media/dirs on stretch, and mounting the wheezy partitions to them for reading purposes. I faintly recall, on my first install attempt, doing that but could only see the partitions root despite trying to cd up a known path, and when I had rebooted to wheezy, the partitions were marked dirty and had to be e2fsck'd which took about 20 minutes with nothing found. Thats not a very confidence inspiring thing.
Since this is a 2 list post, going the emc-developers too since this install media is theirs, my first attempt used the debian 9.4 1st dvd for install media, and networking Just Worked. But lots of other stuff didn't. My 2nd attempt was with the LCNC image of stretch pointed to by the wiki.linuxcnc.org pages, and networking is dead. And my 3rd attempt will use that media again.
That big a change in filesystems deserves a migration tool.
So I'm going to go fix the missus breakfast, then get to it a 3rd time.
On Tuesday 21 August 2018 03:20:43 Dr. Nikolaus Klepp wrote:
Am Dienstag, 21. August 2018 schrieb Gene Heskett:
On Monday 20 August 2018 10:41:23 Dr. Nikolaus Klepp wrote:
Am Montag, 20. August 2018 schrieb Gene Heskett:
[...] These guys make a pretty good installer simply because they know what it needs and doesn't need to run LinuxCNC right. They don't care if it doesn't run on cousin Molly's washing machine, although with some config effort, it can probably run it, better than the programming it came with.
I certainly had no such problems installing tde on top of the kde I originally loaded onto my Dell running a grizzly G0704 mill on a special wheezy install. So I'll do the same and then install tde on that.
Back later with grins or tears.
Thanks Michelle.
Hi Gene!
You should definitly walk the devuan road. I had a hell of a time with machinekit on BBB, till I found out how to get rid of the systemd stuff. When you run stretch + linuxcnc, you'll most likely get bad surprises (aka. latency > 400000ns) with the 4.*-rtpreempt kernel
Nik
Its tears so far Nik. There are so many differences just in the ext4's used that one of them should be renamed, they will NOT cross mount, ext4 disk to ext4 mount.
Hi Gene!
When you cannot mount the ext4 partions from one another, then there is something very wrong. ext4 can be mounted as ext2 and that should alway work - at least, if the drives and filesystems are ok.
Nik
I'll try that (mounting the old wheezy partitions as ext2) on the next install, but first feed the missus.
Gene Heskett composed on 2018-08-20 07:12 (UTC-0400):
I can't mount the stretch drive and edit it, theres enough diffs between the ext4's that it refuses to mount from wheezy.
...
So first, what do I edit or nuke to get rid of the gnome over-ride?
Boot Wheezy, format the partitions the Stretch installer created. Install like Michele wrote, minimal, and, use the existing partitions.
Use existing is the only way I ever install anything. I get exactly what I want that way.
I don't download installation media. I download installation kernel and initrd, put them where existing Grub can find them, create stanza to do so in existing Grub, and net install minimal. Here are example stanzas, one for Grub, other for Grub2:
title Install Debian via HTTP kernel (hd0,2)/boot/debian/linux64 vga=791 --- net.ifnames=0 ipv6.disable=1 netcfg/get_hostname=myhost netcfg/disable_dhcp=true tasks=standard base-installer/install-recommends=false initrd (hd0,2)/boot/debian/initrd64.gz
menuentry "Install Debian via HTTP" { search --no-floppy --label --set=root k25p03res linuxefi /deb9/linux showopts vga=791 --- net.ifnames=0 ipv6.disable=1 netcfg/get_hostname=myhost netcfg/disable_dhcp=true tasks=standard base-installer/install-recommends=false initrdefi /deb9/initrd.gz }
You can type in the same options booting the regular installation media that you already burned. Or you can download the little net boot iso, and burn and boot from it, doing same: http://ftp.us.debian.org/debian/dists/stretch/main/installer-amd64/current/i...
-
Dan Youngquist -
Dr. Nikolaus Klepp -
Felix Miata -
Gene Heskett -
Jimmy Johnson -
John Pisini -
Michele Calgaro -
Mike Bird -
Nick Koretsky -
Pisini, John
-
Slávek Banko -
Stefan Krusche