said Jim via tde-users: | On Sun, Sep 22, 2024 at 18:23 (-0300), Jim via tde-users wrote: | > On Sun, Sep 22, 2024 at 19:56 (+0000), dep via tde-users wrote: | >> Some months ago I asked here if there were a way to log my ISP's | >> downtime, and got a useful answer. Today I've been attempting to | >> refine it a bit, and there is some progress, but not success. | >> | >> Here's what I have: | >> | >> [code] | >> #!/bin/bash | >> #monitor frontier communications downtime | >> ping -i 3 -O -D 1.1.1.1 | while read row | >> do | >> awk '{ sub(/[0-9]{10}/, strftime("%Y-%m-%d %H:%M:%S", | >> substr($0,2,10))) } 1' <<< "$row"| tee ~/frontier_downtime.log | >> done | >> [/code] | >> | >> Most of it is simply converting internet time to human time, because | >> I might have to use it to support a complaint. It runs fine in a | >> terminal, a nice ping and result every three seconds, but only writes | >> the first line, the first ping result, to the logfile. (And yes, ir's | >> probably sloppy.) | > | > I don't think that is true. I think it is writing the most recent | > ping result to the log file. | > | >> What do I have wrong? | > | > You didn't use the '-a' option to tee. So each invocation of tee | > over-writes the file with the most recent line. | > | > I might suggest something like | > | > ping -i 3 -O -D 1.1.1.1 | while read row | > do | > awk '{ sub(/[0-9]{10}/, strftime("%Y-%m-%d %H:%M:%S", | > substr($0,2,10))) } 1' <<< "$row" | > done | tee -a ~/frontier_downtime.log | > | > to avoid calling a new invocation of tee for every ping result. | > | >> Bonus question: any way to keep it running with the terminal closed? | >> I can't even use & to free the terminal; I'd like to start the thing | >> and have it cook merrrily away without a terminal open or even | >> minimized, and use top or killall to dispose of it if that becomes | >> necessary. | >> | >> I'm pretty sure I'm missing something obvious. | > | > I think maybe you are. | > | > touch ~/frontier_downtime.log | > ping -i 3 -O -D 1.1.1.1 | while read row | > do | > awk '{ sub(/[0-9]{10}/, strftime("%Y-%m-%d %H:%M:%S", | > substr($0,2,10))) } 1' <<< "$row" | > done >> ~/frontier_downtime.log & | > | > You should be able to close that terminal after running that without | > anything bad happening. And if you want to see that is going on from | > some terminal, you can use | > tail -f ~/frontier_downtime.log | > in that terminal. | | After I sent that off, it occurred to me that you don't need a while | loop at all... | | These two commands should all be on one line, just in case your mail | reader breaks the lines. | | ping -i 3 -O -D 1.1.1.1 | awk '{ sub(/[0-9]{10}/, strftime("%Y-%m-%d | %H:%M:%S", substr($0,2,10))) } 1' | tee -a ~/frontier_downtime.log | | Have you thought about losing the microseconds? It probably isn't all | that interesting. | | ping -i 3 -O -D 1.1.1.1 | awk '{ sub(/[0-9]{10}.[0-9]{6}/, | strftime("%Y-%m-%d %H:%M:%S", substr($0,2,10))) } 1' | tee -a | ~/frontier_downtime.log Thanks very much. The last one that works here does contain the while. Though getting rid of the microseconds is, yes, something I'd wished for, I didn't know the syntax to do it. I still don't, apparently! -- dep Pictures: http://www.ipernity.com/doc/depscribe/album Column: https://ofb.biz/author/dep/

said Sharon Kimble via tde-users: | Why use a script, does your linux setup have a program called 'uptimes' | (in Debian) which uses the command `uprecords` which gives your longest | uptimes and your cumulative uptimes and your cumulative downtimes. I have "uptime," which renders a one-line summary of how long the computer has been on, the load, etc., but is silent on connections. I have no "uprecords" or "uprecord." I've looked at a variety of tools. Apparently "gping" is no longer maintained; the Gnome tools don't log and don't allow anything except the Linux ping default, which is once per second. I looked at, installed, and tried unsuccessfully to configure "Smokeping," which is like bringing in a battle tank to get rid of a wasp nest. I had hoped that there would be some lightweight utility that would include what I wanted, which is a record of the times my ISP fails to PIS, as proof of their unreliability, chapter and verse. It seems to have turned out that a script does this when nothing else does. (It performed during one of the failures earlier today.) I've thrown it and the log in a little ~/ subdirectory, stuck an icon on it and put it in Kmenu. So one click and it is running, and I can always monitor it by looking at the terminal that's generated. In that it generated more than a megabyte of log per day, I'd like to cook up a way to automate archiving of the existing log, with date, daily. As it is, I can just move the file into an archive subdirectory and rename it at whatever frequency, probably weekly, I'd like. Then, when it's needed, it will be simple to provide a list of times it's been down, to within three seconds. | Does this help? It's good to know -- I was unaware of the utility -- but in this case it doesn't do what I need. Thanks. -- dep Pictures: http://www.ipernity.com/doc/depscribe/album Column: https://ofb.biz/author/dep/

said jacobheinrich--- via tde-users: | Why not use a network monitoring tool like Nagios? Too much tool for the task, don't you think? My goal is a list of ISP downtimes over a period of time, with logs to prove it. -- dep Pictures: http://www.ipernity.com/doc/depscribe/album Column: https://ofb.biz/author/dep/