[OpenAFS] Odd ubik (?) synchronization problem

Russ Allbery rra@stanford.edu
Mon, 26 Apr 2004 19:37:21 -0700


Derrick J Brashear <shadow@dementia.org> writes:

> One of Chaskiel or Jeff figured it out; 31 bit rollover in the tid
> counter. You restarted, it got happy. If we made ubik servers exit
> cleanly at rollover, you'd barely notice.

Sounds like a good solution to me.  In the meantime, sounds like the
vlserver could use a kick every three months or so, if we maintain our
current transaction rate....

-- 
Russ Allbery (rra@stanford.edu)             <http://www.eyrie.org/~eagle/>