|
|
Back to UserFriendly Strip Comments Index
|
Attn Myke | by quilting_kitty | 2006-11-19 12:55:59 |
|
I know your pain. | by Myke | 2006-06-22 07:33:53 |
|
I'm trying to get swatch working on solaris 7 | by Havoc | 2006-06-22 07:46:04 |
|
feed syslog to a more sane machine? (n/t) | by Myke | 2006-06-22 08:02:08 |
|
Heh. It's not a syslogable thing. | by Havoc | 2006-06-22 08:15:03 |
|
do while loop :) | by Myke | 2006-06-22 08:39:06 |
|
Yep - I thought about that | by Havoc | 2006-06-22 10:34:24 |
|
... why does it take 3 minutes? | by Myke | 2006-06-22 11:59:56 |
| Well, we rotate the log monthly, and it is already |
by Havoc |
2006-06-22 15:14:29 |
over 240mb - usually 400+ when we rotate it. It's mostly 2 entries every 5 min, then two others every 20 min or so, stuff like that. Verbose is an understatement. In the past, the db has crapped itself and not stopped the processes... In the log it registers an error, then automatically begins trying to recover, which usually fails, so it recovers again and again until we manually force a proper recovery. While it is trying, the db is quiescent (offline) and all of our systems stop working. This puts a lot of people twiddling their thumbs. Do they call us immediately? no. Of course not. They prefer thumb-twiddling.
I thought about using a simple tail -f <log> |grep <alert words> - but the exec would be a bit problematic. Need it to email/page us differently for different issues. swatch seems to do it all, but getting it to work is proving to be a PITA.
Maybe logsurfer or logwatcher would work better. Already using logwatch on the main logs just to stay on top of what 18 of my servers are doing. We also run Nagios on them.
I'll check and see if Nagios can be extended to the informix logs - that would be the more elegant solution, perhaps. |
|
[ Reply ] |
|
|
[Todays Cartoon Discussion]
[News Index]
|
|
