[Linux-cachefs] Problems with "Readpage failed on backing file"

[gregm at servu.net.au]

Hi David et al,

We are still havin with cachefiles - even after the last few patches
deployed since the 2.6.30 release.

We have 3 servers -

1 x Primary DB/NFS Fileserver
2 x Clustered web boxes wit NFS caching.

We are still getting cachefilesd failing occasionally on both of the web
boxes. I have tried multiple varants of 2.6.30.x and .31

Here is the output from syslog -

Aug 20 00:57:01 dnetwww2 cachefilesd[7046]: Refilling cull table
Aug 20 00:57:01 dnetwww2 cachefilesd[7046]: Scan complete
Aug 20 00:57:19 dnetwww2 Pid: 182, comm: kslowd Not tainted
2.6.31-rc6-git5 #1
Aug 20 00:57:19 dnetwww2 Call Trace:
Aug 20 00:57:19 dnetwww2 [<ffffffff811434e9>] 0xffffffff811434e9
Aug 20 00:57:19 dnetwww2 [<ffffffff812b5d01>] ? 0xffffffff812b5d01
Aug 20 00:57:19 dnetwww2 [<ffffffff81032f86>] 0xffffffff81032f86
Aug 20 00:57:19 dnetwww2 [<ffffffff810331ff>] 0xffffffff810331ff
Aug 20 00:57:19 dnetwww2 [<ffffffff8104f295>] 0xffffffff8104f295
Aug 20 00:57:19 dnetwww2 [<ffffffff81079aea>] 0xffffffff81079aea
Aug 20 00:57:19 dnetwww2 [<ffffffff8110360d>] 0xffffffff8110360d
Aug 20 00:57:19 dnetwww2 [<ffffffff8108c8c1>] ? 0xffffffff8108c8c1
Aug 20 00:57:19 dnetwww2 [<ffffffff8103320a>] ? 0xffffffff8103320a
Aug 20 00:57:19 dnetwww2 [<ffffffff8108ca67>] 0xffffffff8108ca67
Aug 20 00:57:19 dnetwww2 [<ffffffff810b4478>] 0xffffffff810b4478
Aug 20 00:57:19 dnetwww2 [<ffffffff811015b6>] 0xffffffff811015b6
Aug 20 00:57:19 dnetwww2 [<ffffffff810b46f0>] 0xffffffff810b46f0
Aug 20 00:57:19 dnetwww2 [<ffffffff8113f283>] 0xffffffff8113f283
Aug 20 00:57:19 dnetwww2 [<ffffffff81141768>] ? 0xffffffff81141768
Aug 20 00:57:19 dnetwww2 [<ffffffff8113fb8d>] 0xffffffff8113fb8d
Aug 20 00:57:19 dnetwww2 [<ffffffff810f95d3>] 0xffffffff810f95d3
Aug 20 00:57:19 dnetwww2 [<ffffffff81028af4>] ? 0xffffffff81028af4
Aug 20 00:57:19 dnetwww2 [<ffffffff810f9f04>] 0xffffffff810f9f04
Aug 20 00:57:19 dnetwww2 [<ffffffff81078bdc>] 0xffffffff81078bdc
Aug 20 00:57:19 dnetwww2 [<ffffffff81039590>] ? 0xffffffff81039590
Aug 20 00:57:19 dnetwww2 [<ffffffff81032f86>] ? 0xffffffff81032f86
Aug 20 00:57:19 dnetwww2 [<ffffffff8104f2ba>] ? 0xffffffff8104f2ba
Aug 20 00:57:19 dnetwww2 [<ffffffff81078961>] ? 0xffffffff81078961
Aug 20 00:57:19 dnetwww2 [<ffffffff8104ef7a>] 0xffffffff8104ef7a
Aug 20 00:57:19 dnetwww2 [<ffffffff81011b8a>] 0xffffffff81011b8a
Aug 20 00:57:19 dnetwww2 [<ffffffff8104eeeb>] ? 0xffffffff8104eeeb
Aug 20 00:57:19 dnetwww2 [<ffffffff81011b80>] ? 0xffffffff81011b80
Aug 20 00:57:19 dnetwww2 CacheFiles: I/O Error: Readpage failed on backing
file 4000000000000810
Aug 20 00:57:19 dnetwww2 FS-Cache: Cache cachefiles stopped due to I/O error
Aug 20 00:57:31 dnetwww2 CacheFiles: File cache on sdb1 unregistering
Aug 20 00:57:31 dnetwww2 FS-Cache: Withdrawing cache "mycache"
Aug 20 00:57:31 dnetwww2 cachefilesd[7046]: Refilling cull table
Aug 20 00:57:31 dnetwww2 cachefilesd[7046]: Failed to check object's
in-use state: errno 5 (Input/output error)

FYI these are Virtual Machines running under ESX4i on Dell Poweredge R200's


Any feedback would be greatly appreciated.