[Veritas-bu] Looking for a script: a couple of example solutions
2005-11-10 13:19:25
Subject: |
[Veritas-bu] Looking for a script: a couple of example solutions |
From: |
CJManders AT lbl DOT gov (Christopher Jay Manders) |
Date: |
Thu, 10 Nov 2005 10:19:25 -0800 |
Hi,
We use a couple of scripts that might be interesting to someone. I have
rolled up a tar file with them for anyone that wants them. Download from
here: http://backups.lbl.gov/backups/software/error-report.tgz
They will definately need to be customized for your situation, since we
use aux databases (mysql/oracle/ldap) to keep track of the errors in an
ongoing way.
The first script (Daily-Errors.pl) is for the admins. It strips out the
successful and then figures out how long the issue has been going on for
(db hook here). It then groups the errors into two camps: transient and
persistent errors. These are then inserted into an excel spreadsheet via
the CPAN module for that, and then sent to the email address set in the
script with an attachment. Note that we actually put alot of data in
this, so that it can be given to our help desk and have them go thru and
call folks. This uses methods you may not have available...those can be
stripped out, but I left them as is for interest's sake. It also does a
'ping' test, a 'telnet $host 13782' test, and a 'bpdir' test and reports
those per host. We like this script a lot.
The second script (Daily-Email-Notify.pl) actually sends (daily) an
email that has hooks into our Remedy Help Desk system, and also performs
some limited escalation. Eg. for x days (in database per host) notify
x AT lbl DOT gov 1st contact, after the x days, then start to CC the 2nd
contact on file for the host. After x days it then again will
auto-generate a Remedy ticket (i did not include all of this part...if
you want it, then email me offline). At the end of the escalation we
actually take an action (automated and this is also not included here)
of deleting the host using bpclclients, but track it still in our
database with a start and end date. This is monitored daily and hosts
returning from their 'backups vacation' (what we call it) are
re-inserted. This part currently is integrated with our enterprise LDAP
service that has hooks into our Oracle Calendar setup to allow someone
to specify the backup vacation thru the normal calendaring product used
here at LBNL.
I hope these are of use to someone somewhere.
Any comments or feedback welcomed.
Cheers!
--Chris
Jerry wrote:
>bperror -backstat -by_statcode -U
>
>That will show the error code 0 ones too, but you
>could craft a script based on that if you had to.
>
>--- "Major, Rusty" <RMajor AT vericenter DOT com> wrote:
>
>
>
>>Does anyone have a script already written that will
>>poll the jobs db and
>>pull out any jobs that are either failed or not
>>completed and put them
>>in a text file? Just don't want to reinvent the
>>wheel if I don't have
>>to.
>>
>>Thanks,
>>
>>Rusty Major, MCSE, BCFP
>>Data Assurance Engineer
>>(281) 584-4693
>>VeriCenter, Inc.
>>
>>
>>
>>
>
>
>
>
>__________________________________
>Yahoo! FareChase: Search multiple travel sites in one click.
>http://farechase.yahoo.com
>_______________________________________________
>Veritas-bu maillist - Veritas-bu AT mailman.eng.auburn DOT edu
>http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu
>
>
|
|
|