- Wednesday, August 13, 2014

openafs@conference.openafs.org

Wednesday, August 13, 2014< ^ >

Room Configuration

Room Occupants

GMT+0
[08:06:12] deason leaves the room
[11:56:13] meffie joins the room
[12:00:57] wiesand joins the room
[12:46:52] meffie leaves the room
[13:57:01] meffie joins the room
[14:01:54] deason joins the room
[14:02:43] kaduk joins the room
[15:04:17] mvita leaves the room
[15:04:18] mvita joins the room
[16:11:20] meffie leaves the room
[16:15:41] wiesand leaves the room
[17:17:38] meffie joins the room
[17:51:37] kaduk leaves the room
[18:15:16] meffie leaves the room
[20:45:17] kaduk joins the room
[21:16:00] meffie joins the room
[21:38:52] <kaduk> Hmm, have I pestered meffie about afs3-stds I-Ds recently? I think it's been a while...
[21:49:22] <meffie> it has. ok. ok.
[21:50:12] <meffie> (thank you for pestering, i need it:)
[21:50:32] <kaduk> Happy to oblige :)
Let me know if it becomes too frequent.
[21:53:56] <meffie> but first, i wonder if 'afs_cacheStats' in afs_FlushVCBs is really supposed to be 'afs_vcount'?
[21:55:28] <meffie> (or afs_maxvcount)
[22:06:33] <kaduk> It looks like the comparison is against a number of fids observed. afs_cacheStats claims to be a number of stat entries in cache.
[22:07:16] <kaduk> But, afs_cacheStats does not appear to change after init.
[22:07:21] <Jeffrey Altman> The use of afs_cacheStats in afs_FlushVCBs was introduced between afs 3.3 and 3.4 and bounded the number of callbacks broken based upon the number of stat entries.
[22:07:22] <meffie> well, on linux it's the starting number.
[22:08:37] <Jeffrey Altman> the bound was added to protect against infinite loops in the list due to corruption of the list
[22:10:36] <Jeffrey Altman> it might very well be the case that (a) the list corruption no longer occurs and the check is not required; and (b) that the assumptions that were true prior to the release of 3.4 no longer hold true
[22:14:36] <meffie> ok, thanks. i just noticed it, and was wondering out loud if it really should be "afs_vcount", since that can exceed afs_cacheStats after the dynamic vcache patch was committed.
[22:16:56] <meffie> but if it's a limit, i suppose it's ok to be less.
[22:18:42] kaduk leaves the room
[22:22:18] meffie leaves the room
[23:05:15] meffie joins the room
[23:05:19] meffie leaves the room
[23:27:47] deason leaves the room