Opened 9 months ago

Last modified 6 months ago

#31784 assigned task

move majus off of moly

Reported by: anarcat Owned by: hiro
Priority: Medium Milestone:
Component: Internal Services/Tor Sysadmin Team Version:
Severity: Normal Keywords: l10n
Cc: emmapeel Actual Points:
Parent ID: #29974 Points:
Reviewer: Sponsor:

Description

as part of an effort to reduce our dependence on an old server (moly), we should move majus (a non-redundant server) to a different host, probably the FSN* cluster.

Child Tickets

Change History (4)

comment:1 Changed 8 months ago by hiro

Owner: changed from tpa to hiro
Status: newassigned

comment:2 Changed 8 months ago by emmapeel

Cc: emmapeel added
Keywords: l10n added

comment:3 Changed 6 months ago by anarcat

this might need to be done urgently after all, because moly is having serious networking issues (in #32762), stay tuned.

comment:4 Changed 6 months ago by anarcat

I may need a little help on this. Where are the backups or where can i see what is being backed up, to spot problems?

That's a good question! Backups are done on our central backup server, using the "Standardbackup" job, which uses the "Standard set", which looks like this:

FileSet {
  Name = "Standard Set"
  Ignore FileSet Changes = yes
  Include {
    Options {
      signature = SHA1
      compression = GZIP9
      aclsupport = yes
      xattrsupport = yes
    }
    Options {
      wild = "/swapfile*"
      exclude = yes
    }
    File = "\\|/usr/local/sbin/bacula-backup-dirs"
    # Dont backup directories that contain .nobackup files
    Exclude Dir Containing = .nobackup
  }

  Exclude {
    File = /.fsck
    File = /.journal
    File = /dev
    File = /home/buildd/build-trees
    File = /lib/init/rw
    File = /nonexistant
    File = /proc
    File = /srv/chroot
    File = /sys
    File = /tmp
    File = /srv/apache-cache/mod_cache_disk
    File = /var/cache/apache2/mod_cache_disk
    File = /var/cache/apt
    File = /var/lib/apt
    File = /var/lib/bacula
    File = /var/lib/munin-async
    File = /var/lock
    File = /var/log/samhain
    File = /var/run
    File = "\\|bash -c 'grep -s -v ^# /etc/bacula/local-exclude || true'"
  }
}

That file parameter tells Bacula to call a script to see which files to backup, and that script on majus returns:

root@majus:~# /usr/local/sbin/bacula-backup-dirs
/
/boot
/srv

So basically everything backed up on this server, with the notable exception of that Exclude list.

or shall I may a list of what I think should be moved?

That might be useful, I'm not sure. Maybe it's best if hiro or I just copy everything to the new server, unless you see a reason why we should make it different?

Also, are we going to have some puppet thing running the new server?

I'm not sure what you're asking here... Like all TPA machines, we do have Puppet running on this server and it will also run on the next one. We don't, however, have any *specific* configuration for translation in Puppet for Majus, so everything is configured by hand on this server.

I would be happy to integrate more Puppet stuff on this server, exactly to work around issues like the ones we faced this week. If more components of the service were in Puppet, it would be easier to restore the service: just run Puppet to recreate the service and restore only the data, as opposed to doing a barebones backup restore which is a more complex operation.

I hope that answers your questions!

Note: See TracTickets for help on using tickets.