upgrade process?

Colin Surprenant colin.surprenant at gmail.com
Tue Nov 23 10:58:19 EST 2010


On Tue, Nov 23, 2010 at 10:18 AM, David Smith <dizzyd at basho.com> wrote:
> In our internal testing we have started verifying that
> sequential releases (i.e. 0.12 -> 0.13) support the rolling upgrades
> cleanly.

Great stuff. May I suggest to add this fact into the rolling upgrade wiki page?

> Given the leap of the upgrade you are describing, I would _strongly_
> recommend setting up a test environment that's as similar to your
> production as possible (load, data set, etc). As I noted in my
> previous email, it should work, but I can't guarantee it. :)

Would upgrading (rolling or offline) sequentially from 0.10 to 0.11 to
0.12 be "easier" to manage potential compatibility issues?

> I'd note that if you're using bitcask, you can do such a backup
> without ever stopping the server. Innostore is less straightforward,
> unfortunately.

I'm using innostore, with a 3 nodes cluster holding ~250GB of data
each. Using the riak-admin backup/restore is just too slow to be
usable in my environment. Would innodump/innoload be any faster? Could
I consider setting up a new 0.13 ring, use innodump on my 0.10 nodes
and innoload on the 0.13??

> My vagueness re: 0.10 -> 0.13 was driven by the simple reality that we
> haven't tested it formally and we don't have a policy of testing an
> upgrade across that many versions.

I understand. The last time I checked, Riak was still a version 0.xx
right? ;) Seriously, It's great news that you guys are taking better
care of the upgrade process starting at 0.12.

Thanks for your support,
Colin




More information about the riak-users mailing list