About Archive Tags RSS Feed

 

Getting ready for Stretch

25 May 2017 21:50

I run about 17 servers. Of those about six are very personal and the rest are a small cluster which are used for a single website. (Partly because the code is old and in some ways a bit badly designed, partly because "clustering!", "high availability!", "learning!", "fun!" - seriously I had a lot of fun putting together a fault-tolerant deployment with haproxy, ucarp, etc, etc. If I were paying for it the site would be both retired and static!)

I've started the process of upgrading to stretch by picking a bunch of hosts that do things I could live without for a few days - in case there were big problems, or I needed to restore from backups.

So far I've upgraded:

  • master.steve
    • This is a puppet-master, so while it is important killing it wouldn't be too bad - after all my nodes are currently setup properly, right?
    • Upgrading this host changed the puppet-server from 3.x to 4.x.
    • That meant I had to upgrade all my client-systems, because puppet 3.x won't talk to a 4.x master.
    • Happily jessie-backports contains a recent puppet-client.
    • It also meant I had to rework a lot of my recipes, in small ways.
  • builder.steve
    • This is a host I use to build packages upon, via pbuilder.
    • I have chroots setup for wheezy, jessie, and stretch, each in i386 and amd64 flavours.
  • git.steve
    • This is a host which stores my git-repositories, via gitbucket.
    • While it is an important host in terms of functionality, the software it needs is very basic: nginx proxies to a java application which runs on localhost:XXXX, with some caching magic happening to deal with abusive clients.
    • I do keep considering using gitlab, because I like its runners, etc. But that is pretty resource intensive.
    • On the other hand If I did switch I could drop my builder.steve host, which might mean I'd come out ahead in terms of used resources.
  • leave.steve
    • Torrent-box.
    • Upgrading was painless, I only run rtorrent, and a simple object storage system of my own devising.

All upgrades were painless, with only one real surprise - the attic-backup software was removed from Debian.

Although I do intend to retry using Larss' excellent obnum in the near future pragmatically I wanted to stick with what I'm familiar with. Borg backup is a fork of attic I've been aware of for a long time, but I never quite had a reason to try it out. Setting it up pretty much just meant editing my backup-script:

s/attic/borg/g

Once I did that, and created some new destinations all was good:

[email protected] ~ $ borg init /backups/git.steve.org.uk.borg/
[email protected] ~ $ borg init /backups/master.steve.org.uk.borg/
[email protected] ~ $ ..

Upgrading other hosts, for example my website(s), and my email-box, will be more complex and fiddly. On that basis they will definitely wait for the formal stretch release.

But having a couple of hosts running the frozen distribution is good for testing, and to let me see what is new.

| 1 comment

 

Comments on this entry

icon Tomasz at 11:15 on 28 May 2017

You could virtualize all of them on one or two boxes for failover, and then you could easily duplicate each one to test the upgrades. This would keep the existing infra running while you could prep the new ones for swapover.