Project

General

Profile

Bug #9074

dCache testbed depends on chimeradump-disk-0.9-5 from the disk instance.

Added by Natalia Ratnikova over 4 years ago. Updated over 4 years ago.

Status:
Assigned
Priority:
Normal
Start date:
06/08/2015
Due date:
% Done:

0%

Estimated time:
Spent time:
component:
base
First Occurred:
Occurs In:
Stakeholders:
Co-Assignees:
Duration:

Description

The scripts contained in the rpm refer to the disk instance servers,
The rpm should be rebuilt, or disabled.

History

#1 Updated by Natalia Ratnikova over 4 years ago

  • Status changed from New to Assigned
  • Assignee set to Natalia Ratnikova
  • Priority changed from Normal to High

Increase the priority, as there is a cron job running on the testbed,
which acts on the production disk instance:

[root@cmspnfs1 chimera-list]# for s in `cat /tmp/cd_files.NR`; do grep $s /tmp/crontab.NR

done

/10 * * * * /storage/local/data1/copyOnceWalls-disk.sh >/dev/null 2>&1
/15 * * * * /usr/libexec/dcache/checkPostgres.sh >/dev/null 2>&1
10 7 * * * /usr/libexec/dcache/createTARbackup.sh >/dev/null 2>&1
/15 * * * * /usr/libexec/dcache/pnfsManager.monitor >/dev/null 2>&1
/15 * * * * /usr/libexec/dcache/poolManager.monitor >/dev/null 2>&1
*/30 * * * * /usr/libexec/dcache/poolsEnabled.monitor >/dev/null 2>&1

Even if it does not break any things, it creates extra load.

#2 Updated by Natalia Ratnikova over 4 years ago

As a quick fix - remove all cron job entries referring to production disk instance.
They come from commit 68877bcf on Dec 15th, 2014.

Check the contents of the scripts running in the crontab.
The following two rely on certain configuration, but are not instance specific:

/usr/libexec/dcache/checkPostgres.sh
/usr/libexec/dcache/createTARbackup.sh

The following four scripts either use connectChimera-disk.sh script or rsync to cmschimeradisk machine:

/usr/libexec/dcache/poolManager.monitor
/usr/libexec/dcache/pnfsManager.monitor
/usr/libexec/dcache/poolsEnabled.monitor
/storage/local/data1/copyOnceWalls-disk.sh

Created branch:
natalia_redmine_9074.

Crontab is deleted in order to get rid of the removed entries, here are the commands:

service puppet stop
puppet agent -t --environment=natalia_redmine_9074 --noop
puppet agent -t --environment=natalia_redmine_9074
crontab -l > crontab.save.NR
crontab -r
puppet agent -t --environment=natalia_redmine_9074
crontab -l > crontab.new.NR
  1. push cmspnfs1 to natalia_redmine_9074 branch in the ENC
    service puppet start

#3 Updated by Natalia Ratnikova over 4 years ago

The proper solution will be :

1) move configurable items into puppet/hiera
2) package external tools for chimera dumps and local monitoring scripts into separate rpms

Here are some hints on managing DCSO rpms:

https://cmsweb.fnal.gov/bin/view/ComputingServices/Rpm

#4 Updated by Natalia Ratnikova over 4 years ago

  • Priority changed from High to Normal

Lower the priority, as the work around has been applied.



Also available in: Atom PDF