[OpenAFS] 1.2.13 vs. 1.3.81

[Derrick J Brashear]

On Thu, 7 Apr 2005, Marcus Watts wrote:

> We currently run 1.2.13 in production on the fileservers for the
> umich.edu cell, but have now run up against two problems related to
> callback logic of varying severity, which turn out to have fixes in
> recent 1.3.x.  Our production staff is highly allergic to the idea of
> running anything that has the description "unstable", like "unstable
> development branch"; but I'm beginning to think that this is way
> preferable to trying to backport fixes out of recent 1.3 to the 1.2
> branch.
>
> As it happens, we're having a meeting in less than 12 hours to discuss
> all this, so I'm highly curious how people feel on this list: is
> 1.3.80/81 more stable than 1.2.13?  What do people actually run in
> production?  Is there in fact any more work planned for the 1.2
> branch?  Do you have any particular words of advice for our production
> staff if they should be keen on continuing with 1.2.13 despite the
> problems we've had?  Any thoughts on 1.3.80 vs. 1.3.81 vs. whatever?
> (This will all be on linux / x86 hardware.)

A 1.2.14 with known existing fixes (hence needing less testing) is a very 
real possibility, but 1.4 is realistically near also. Perhaps you wish to 
exercise the paranoia with which new software releases are often met, and 
not run the .0 version of it. Perhaps not.

I will let other people comment on what they think of 1.3.8x on their 
machines, but as far as servers I don't think anything which went into .81 
is that exciting.