Slavek,
I doubt you have had time to work on user session tracking in the pure systemd case. However, this is an issue I must get fixed or TDE is unusable on Arch (unless we ask users to manually kill all process that don't close)
Is there anything I can do to help? We have the kdebase-workspace patch from 4.11, have you made an attempt to apply it? If not, I suppose I could hand-edit/apply as much as I can, and then provide you with that if you think it will help.
Let me know if that would help or if there is something else I can do to help move this along. Thanks.
On Tuesday 11 of February 2014 02:14:52 David C. Rankin wrote:
Slavek,
I doubt you have had time to work on user session tracking in the pure systemd case. However, this is an issue I must get fixed or TDE is unusable on Arch (unless we ask users to manually kill all process that don't close)
Is there anything I can do to help? We have the kdebase-workspace patch from 4.11, have you made an attempt to apply it? If not, I suppose I could hand-edit/apply as much as I can, and then provide you with that if you think it will help.
Let me know if that would help or if there is something else I can do to help move this along. Thanks.
Currently I have unfinished a patch for kdesktop_lock => response to systemd signals Lock(), Unlock() and setting SetIdleHint(true / false). I'll put patch to test when will be ready.
Then I want to prepare minimal systemd support for TDM.
And as a last I want to deal with automatic MultiSeat with systemd. But I would like to use D-Bus calls instead of a direct call API functions sd_*.
On 02/10/2014 08:14 PM, Slávek Banko wrote:
Currently I have unfinished a patch for kdesktop_lock => response to systemd signals Lock(), Unlock() and setting SetIdleHint(true / false). I'll put patch to test when will be ready.
Then I want to prepare minimal systemd support for TDM.
And as a last I want to deal with automatic MultiSeat with systemd. But I would like to use D-Bus calls instead of a direct call API functions sd_*.
Excellent. Just let me know when you want me to test. Thanks.
On 02/10/2014 09:55 PM, David C. Rankin wrote:
On 02/10/2014 08:14 PM, Slávek Banko wrote:
Currently I have unfinished a patch for kdesktop_lock => response to systemd signals Lock(), Unlock() and setting SetIdleHint(true / false). I'll put patch to test when will be ready.
Then I want to prepare minimal systemd support for TDM.
And as a last I want to deal with automatic MultiSeat with systemd. But I would like to use D-Bus calls instead of a direct call API functions sd_*.
Excellent. Just let me know when you want me to test. Thanks.
Slavek,
We may need to bump priority on this one - it is worse than initially reported. It doesn't just effect fringe tdeio_slaves like the sftp tdeio_slave it effects *ALL* tdeio_slaves, including tdeio_http. I browsed in konqueror looking for configuration help with Tork and for help with the mixminion python error and I had 50+ tdeio_file and tdeio_http processes like the following:
... tdeio_file [tdeinit] file /tmp/tdesocket-david/tdelauncher0fCUxj.s tdeio_file [tdeinit] file /tmp/tdesocket-david/tdelauncher0fCUxj.s tdeio_file [tdeinit] file /tmp/tdesocket-david/tdelauncher0fCUxj.s tdeio_http [tdeinit] https /tmp/tdesocket-david/tdelauncher0fCUxj. tdeio_http [tdeinit] https /tmp/tdesocket-david/tdelauncher0fCUxj. ...
Apparently with every http connection, including browser search, a tdeio_file slave is also opened to check/update the cache entries in addition to the tdeio_http slave opened for the connection, so this will spin out of control quickly with internet browsing.
On Thursday 13 of February 2014 01:48:38 David C. Rankin wrote:
On 02/10/2014 09:55 PM, David C. Rankin wrote:
On 02/10/2014 08:14 PM, Slávek Banko wrote:
Currently I have unfinished a patch for kdesktop_lock => response to systemd signals Lock(), Unlock() and setting SetIdleHint(true / false). I'll put patch to test when will be ready.
Then I want to prepare minimal systemd support for TDM.
And as a last I want to deal with automatic MultiSeat with systemd. But I would like to use D-Bus calls instead of a direct call API functions sd_*.
Excellent. Just let me know when you want me to test. Thanks.
Slavek,
We may need to bump priority on this one - it is worse than initially reported. It doesn't just effect fringe tdeio_slaves like the sftp tdeio_slave it effects *ALL* tdeio_slaves, including tdeio_http. I browsed in konqueror looking for configuration help with Tork and for help with the mixminion python error and I had 50+ tdeio_file and tdeio_http processes like the following:
... tdeio_file [tdeinit] file /tmp/tdesocket-david/tdelauncher0fCUxj.s tdeio_file [tdeinit] file /tmp/tdesocket-david/tdelauncher0fCUxj.s tdeio_file [tdeinit] file /tmp/tdesocket-david/tdelauncher0fCUxj.s tdeio_http [tdeinit] https /tmp/tdesocket-david/tdelauncher0fCUxj. tdeio_http [tdeinit] https /tmp/tdesocket-david/tdelauncher0fCUxj. ...
Apparently with every http connection, including browser search, a tdeio_file slave is also opened to check/update the cache entries in addition to the tdeio_http slave opened for the connection, so this will spin out of control quickly with internet browsing.
I am not sure whether this behavior is due to session management. That is very reminds me the bug report 1656.
http://bugs.trinitydesktop.org/show_bug.cgi?id=1656
However, it does not change the fact that I'll continue to work to supplement sytemd support. Unfortunately, I have in recent days been too busy with my work tasks.
Slavek --
On Thursday 13 of February 2014 03:13:02 Slávek Banko wrote:
On Thursday 13 of February 2014 01:48:38 David C. Rankin wrote:
On 02/10/2014 09:55 PM, David C. Rankin wrote:
On 02/10/2014 08:14 PM, Slávek Banko wrote:
Currently I have unfinished a patch for kdesktop_lock => response to systemd signals Lock(), Unlock() and setting SetIdleHint(true / false). I'll put patch to test when will be ready.
Then I want to prepare minimal systemd support for TDM.
And as a last I want to deal with automatic MultiSeat with systemd. But I would like to use D-Bus calls instead of a direct call API functions sd_*.
Excellent. Just let me know when you want me to test. Thanks.
Slavek,
We may need to bump priority on this one - it is worse than initially reported. It doesn't just effect fringe tdeio_slaves like the sftp tdeio_slave it effects *ALL* tdeio_slaves, including tdeio_http. I browsed in konqueror looking for configuration help with Tork and for help with the mixminion python error and I had 50+ tdeio_file and tdeio_http processes like the following:
... tdeio_file [tdeinit] file /tmp/tdesocket-david/tdelauncher0fCUxj.s tdeio_file [tdeinit] file /tmp/tdesocket-david/tdelauncher0fCUxj.s tdeio_file [tdeinit] file /tmp/tdesocket-david/tdelauncher0fCUxj.s tdeio_http [tdeinit] https /tmp/tdesocket-david/tdelauncher0fCUxj. tdeio_http [tdeinit] https /tmp/tdesocket-david/tdelauncher0fCUxj. ...
Apparently with every http connection, including browser search, a tdeio_file slave is also opened to check/update the cache entries in addition to the tdeio_http slave opened for the connection, so this will spin out of control quickly with internet browsing.
I am not sure whether this behavior is due to session management. That is very reminds me the bug report 1656.
http://bugs.trinitydesktop.org/show_bug.cgi?id=1656
However, it does not change the fact that I'll continue to work to supplement sytemd support. Unfortunately, I have in recent days been too busy with my work tasks.
Slavek
The first part for the test:
Into kdesktop is added dBus connection and response to SystemD signals Lock() and Unlock() - true, on Unlock() is nothing performed. At the same time on activate / deactivate the screen saver is called SetIdleHint(true / false).
Slavek --
On 02/12/2014 10:52 PM, Slávek Banko wrote:
On Thursday 13 of February 2014 03:13:02 Slávek Banko wrote:
On Thursday 13 of February 2014 01:48:38 David C. Rankin wrote:
On 02/10/2014 09:55 PM, David C. Rankin wrote:
On 02/10/2014 08:14 PM, Slávek Banko wrote:
Currently I have unfinished a patch for kdesktop_lock => response to systemd signals Lock(), Unlock() and setting SetIdleHint(true / false). I'll put patch to test when will be ready.
Then I want to prepare minimal systemd support for TDM.
And as a last I want to deal with automatic MultiSeat with systemd. But I would like to use D-Bus calls instead of a direct call API functions sd_*.
Excellent. Just let me know when you want me to test. Thanks.
Slavek,
We may need to bump priority on this one - it is worse than initially reported. It doesn't just effect fringe tdeio_slaves like the sftp tdeio_slave it effects *ALL* tdeio_slaves, including tdeio_http. I browsed in konqueror looking for configuration help with Tork and for help with the mixminion python error and I had 50+ tdeio_file and tdeio_http processes like the following:
... tdeio_file [tdeinit] file /tmp/tdesocket-david/tdelauncher0fCUxj.s tdeio_file [tdeinit] file /tmp/tdesocket-david/tdelauncher0fCUxj.s tdeio_file [tdeinit] file /tmp/tdesocket-david/tdelauncher0fCUxj.s tdeio_http [tdeinit] https /tmp/tdesocket-david/tdelauncher0fCUxj. tdeio_http [tdeinit] https /tmp/tdesocket-david/tdelauncher0fCUxj. ...
Apparently with every http connection, including browser search, a tdeio_file slave is also opened to check/update the cache entries in addition to the tdeio_http slave opened for the connection, so this will spin out of control quickly with internet browsing.
I am not sure whether this behavior is due to session management. That is very reminds me the bug report 1656.
http://bugs.trinitydesktop.org/show_bug.cgi?id=1656
However, it does not change the fact that I'll continue to work to supplement sytemd support. Unfortunately, I have in recent days been too busy with my work tasks.
Slavek
The first part for the test:
Into kdesktop is added dBus connection and response to SystemD signals Lock() and Unlock() - true, on Unlock() is nothing performed. At the same time on activate / deactivate the screen saver is called SetIdleHint(true / false).
Slavek
Excellent, I'll rebuild and report back.
On 02/12/2014 10:52 PM, Slávek Banko wrote:
On Thursday 13 of February 2014 03:13:02 Slávek Banko wrote:
On Thursday 13 of February 2014 01:48:38 David C. Rankin wrote:
On 02/10/2014 09:55 PM, David C. Rankin wrote:
On 02/10/2014 08:14 PM, Slávek Banko wrote:
Currently I have unfinished a patch for kdesktop_lock => response to systemd signals Lock(), Unlock() and setting SetIdleHint(true / false). I'll put patch to test when will be ready.
Then I want to prepare minimal systemd support for TDM.
And as a last I want to deal with automatic MultiSeat with systemd. But I would like to use D-Bus calls instead of a direct call API functions sd_*.
Excellent. Just let me know when you want me to test. Thanks.
Slavek,
We may need to bump priority on this one - it is worse than initially reported. It doesn't just effect fringe tdeio_slaves like the sftp tdeio_slave it effects *ALL* tdeio_slaves, including tdeio_http. I browsed in konqueror looking for configuration help with Tork and for help with the mixminion python error and I had 50+ tdeio_file and tdeio_http processes like the following:
... tdeio_file [tdeinit] file /tmp/tdesocket-david/tdelauncher0fCUxj.s tdeio_file [tdeinit] file /tmp/tdesocket-david/tdelauncher0fCUxj.s tdeio_file [tdeinit] file /tmp/tdesocket-david/tdelauncher0fCUxj.s tdeio_http [tdeinit] https /tmp/tdesocket-david/tdelauncher0fCUxj. tdeio_http [tdeinit] https /tmp/tdesocket-david/tdelauncher0fCUxj. ...
Apparently with every http connection, including browser search, a tdeio_file slave is also opened to check/update the cache entries in addition to the tdeio_http slave opened for the connection, so this will spin out of control quickly with internet browsing.
I am not sure whether this behavior is due to session management. That is very reminds me the bug report 1656.
http://bugs.trinitydesktop.org/show_bug.cgi?id=1656
However, it does not change the fact that I'll continue to work to supplement sytemd support. Unfortunately, I have in recent days been too busy with my work tasks.
Slavek
The first part for the test:
Into kdesktop is added dBus connection and response to SystemD signals Lock() and Unlock() - true, on Unlock() is nothing performed. At the same time on activate / deactivate the screen saver is called SetIdleHint(true / false).
Slavek
Slavek,
I rebuilt and installed tdebase with the patch. So far, it is not closing any of the tdeio-slave processes. I do get new-looking gpg processes started, but it isn't closing anything. I've saved the xsession-errors file and the running process list as it existed after I used sftp in konqueror to open my remote tde folders. (they are attached) Looking at xsession-errors, I see the normal failed to close:
tdeio_sftp: ERROR: KSshProcess::version(): pclose failed.
but now it looks like I also get:
tdeio (TDEIOConnection): ERROR: Could not write data (pid 6536 process "it] sftp /tmp/tdesocket-david/tdelauncher5rc4DS.s") tdeio_sftp: ERROR: KSshProcess::version(): pclose failed. tdeio (TDEIOConnection): ERROR: Could not write data (pid 6536 process "it] sftp /tmp/tdesocket-david/tdelauncher5rc4DS.s") tdeio (TDEIOConnection): ERROR: Could not write data (pid 6536 process "it] sftp /tmp/tdesocket-david/tdelauncher5rc4DS.s") tdeio_sftp: ERROR: sftpStat(): stat failed with code 2 tdeio_sftp: ERROR: sftpStat(): stat failed with code 2
So it looks like we are making progress and trying to effect closure, we just are not there yet. Here is what I see new in the running process list:
6520 ? S 0:00 _ tdeio_sftp [tdeinit] sftp /tmp/tdesocket-david/tdelauncher5rc4DS.s 6523 pts/3 Ss+ 0:00 | _ /usr/bin/ssh -o ForwardX11=no -o ForwardAgent=no -e none -l david -v phoinix.rlfpllc.com -s sftp
Let me know what else you need and when you have another patch to test. So far so good -- we have new output!
-
David C. Rankin -
Slávek Banko