So what I need to accomplish is to make a unix shell script that I assume I will run as a cron task. This script will need to do an action based on if it sees an entry x number of times in y period of time. It needs to read a file that will look like this
Nov 13, 2008 4:18:22 PM CET NBBS Major NBBS-6000 Cluster node dropped: 10.97.39.208:32984
Nov 13, 2008 4:18:36 PM CET NBBS Normal NBBS-6001 Cluster node joined: 10.97.39.208:32990
Nov 13, 2008 4:18:50 PM CET NBBS Major NBBS-6000 Cluster node dropped: 10.97.39.208:32990
Nov 13, 2008 4:19:04 PM CET NBBS Normal NBBS-6001 Cluster node joined: 10.97.39.208:32996
Nov 13, 2008 4:19:18 PM CET NBBS Major NBBS-6000 Cluster node dropped: 10.97.39.208:32996
Nov 13, 2008 4:19:32 PM CET NBBS Normal NBBS-6001 Cluster node joined: 10.97.39.208:33002
Nov 13, 2008 4:19:46 PM CET NBBS Major NBBS-6000 Cluster node dropped: 10.97.39.208:33002
Nov 13, 2008 4:19:58 PM CET NBBS Normal NBBS-6001 Cluster node joined: 10.97.39.208:33008
Nov 13, 2008 4:20:11 PM CET NBBS Major NBBS-6000 Cluster node dropped: 10.97.39.208:33008
Nov 13, 2008 4:20:25 PM CET NBBS Normal NBBS-6001 Cluster node joined: 10.97.39.208:33014
Nov 13, 2008 4:20:39 PM CET NBBS Major NBBS-6000 Cluster node dropped: 10.97.39.208:33014
Nov 13, 2008 4:20:52 PM CET NBBS Normal NBBS-6001 Cluster node joined: 10.97.39.208:33020
Nov 13, 2008 4:21:06 PM CET NBBS Major NBBS-6000 Cluster node dropped: 10.97.39.208:33020
Nov 13, 2008 4:21:19 PM CET NBBS Normal NBBS-6001 Cluster node joined: 10.97.39.208:33026
Nov 13, 2008 4:21:32 PM CET NBBS Major NBBS-6000 Cluster node dropped: 10.97.39.208:33026
Nov 13, 2008 4:21:46 PM CET NBBS Normal NBBS-6001 Cluster node joined: 10.97.39.208:33032
Nov 13, 2008 4:21:59 PM CET NBBS Major NBBS-6000 Cluster node dropped: 10.97.39.208:33032
Nov 13, 2008 4:22:15 PM CET NBBS Normal NBBS-6001 Cluster node joined: 10.97.39.208:33038
Nov 13, 2008 4:22:29 PM CET NBBS Major NBBS-6000 Cluster node dropped: 10.97.39.208:33038
Nov 13, 2008 4:22:43 PM CET NBBS Normal NBBS-6001 Cluster node joined: 10.97.39.208:33044
Nov 13, 2008 4:22:56 PM CET NBBS Major NBBS-6000 Cluster node dropped: 10.97.39.208:33044
Nov 13, 2008 4:23:10 PM CET NBBS Normal NBBS-6001 Cluster node joined: 10.97.39.208:33050
Nov 13, 2008 4:23:24 PM CET NBBS Major NBBS-6000 Cluster node dropped: 10.97.39.208:33050
Nov 13, 2008 4:23:36 PM CET NBBS Normal NBBS-6001 Cluster node joined: 10.97.39.208:33056
Nov 13, 2008 4:23:49 PM CET NBBS Major NBBS-6000 Cluster node dropped: 10.97.39.208:33056
Nov 13, 2008 4:24:02 PM CET NBBS Normal NBBS-6001 Cluster node joined: 10.97.39.208:33062
Nov 13, 2008 4:24:17 PM CET NBBS Major NBBS-6000 Cluster node dropped: 10.97.39.208:33062
Nov 13, 2008 4:41:56 PM CET NBBS Normal NBBS-6001 Cluster node joined: 10.97.39.208:32782
Nov 14, 2008 1:26:50 AM CET NBBS Major NBBS-6000 Cluster node dropped: 10.97.39.193:33067
Nov 14, 2008 1:26:53 AM CET NBBS Normal NBBS-6001 Cluster node joined: 10.97.39.193:33073
Nov 14, 2008 8:54:36 PM CET NBBS Major NBBS-6000 Cluster node dropped: 10.97.39.193:33073
Nov 14, 2008 8:54:58 PM CET NBBS Normal NBBS-6001 Cluster node joined: 10.97.39.193:33079
Nov 15, 2008 4:42:27 PM CET NBBS Critical NBBS-2000 Cannot contact database
Nov 15, 2008 4:42:27 PM CET NBBS Critical NBBS-2000 Cannot contact database
Nov 15, 2008 4:42:27 PM CET NBBS Critical NBBS-2000 Cannot contact database
I will need to look for event Cluster node dropped and have a setable x and y value. for example if x was 5 and y was 30 I will need to figure out if that line has appeared 5 times in the last 30 minutes. I tried converting things to epoch time and comparing but not getting very far. I also had something that would work in perl but perl is removed from these machines (Solaris 9 sparc) and cannot be installed.
Bash is available
Also if anyone has any ideas as to how I can tell if I already restarted and if the events have already been seen (not a good explaination) that would be nice too.
At the end of the day If a node has dropped x time in y period of time I want to restart the node but then I don't want to restart it again unless there are x new drops in y period. I also need to get the ip out of the message for the node this happened on to restart. thanks in advance.
Start Free Trial
