Failed to read cache clean timestamp - why?

[Ovidiu]

This is the relevant log part:

[Tue Apr 03 09:05:40 2012] [warn] [mod_pagespeed 0.10.21.2-1381 @22777] Failed to read cache clean timestamp /var/cache/mod_pagespeed/!clean!time!.  Doing an extra cache clean to be safe.

checked the permissions:

h1870666:~# ls -al /var/cache/mod_pagespeed/

total 8

drwxrwxr-x  4 www-data www-data  100 Apr  3 09:05 .

drwxr-xr-x 19 root     root     4096 Apr  3 08:51 ..

-rw-r--r--  1 www-data www-data   13 Apr  3 09:05 !clean!time!

drwxr-xr-x 10 www-data www-data  200 Apr  3 09:01 http,3A

drwxr-xr-x 10 www-data www-data  200 Apr  3 08:57 rname

h1870666:~# 

Looks alright to me so whats wrong here?

[Joshua Marantz]

If these are happening infrequently (say once an hour or less) then I wouldn't worry about it.  The way that cache-cleaning works is that every Apache child process wakes up after the cache-clean interval & simultaneously attempts to clean the cache, checking that timestamp file.  If the server has a fair amount of traffic then two child processes will race and the loser will print this message and there will be some extra effort expended redundantly scanning the file-cache for exceeded limits.  It doesn't actually mean anything is wrong, and perhaps we should not warn about it.

-Josh

[Ovidiu]

It happened indeed infrequently.
Thanks for clarifying.