Means that you have too few directory entries for that disk. Odd, unless you are caching very small objects or have set the configuration for average object size too high.
john On Wed, May 2, 2012 at 7:37 AM, Bruce Lysik <[email protected]> wrote: > Great, thanks for confirming that. Back to the initial set of messages > for a moment: > > [May 1 09:23:11.564] Server {0x2afe89312700} WARNING: cache directory > overflow on '/dev/sdd' segment 0, purging... > [May 1 09:23:11.565] Server {0x2afe89312700} WARNING: cache directory > overflow on '/dev/sdd' segment 0, purging... > > This sounds like something it would eventually recover from, but we > haven't seen these messages ever stop, once they do. Can you shine some > light on what's happening here? ATS is functioning properly in this > situation, but load increases, and eventually more cache directories get > into this state. > > Sounds like something I should file a bug against? > > Thanks for your assistance. > > -- > Bruce Z. Lysik <[email protected]> > > ------------------------------ > *From:* Leif Hedstrom <[email protected]> > *To:* [email protected] > *Cc:* Bruce Lysik <[email protected]>; John Plevyak <[email protected]> > *Sent:* Wednesday, May 2, 2012 7:26 AM > > *Subject:* Re: selectively clear device disk cache? > > On 5/2/12 7:30 AM, Bruce Lysik wrote: > > Thanks for that suggestion, John. Objects aren't stored across cache > devices, right? If I clear a single disk cache, it has its own index, and > I don't run the risk of putting the ATS server into some weird state, right? > > > > Correct. It's a strength, and a weakness. In some cases, a very large, > popular file could benefit from spanning multiple disks, but we (currently) > don't support that (afaik). > > -- leif > > > >
