Looks like you have a log file reaching the 2GB limit of standard filesystem
interfaces on 32 bit processors.. Most certainly it is the access.log file
which has grown larger than your CPU architecture can handle cleanly.
Recommendation: set up log rotation to have your logs periodically rotated
and cleaned to keep them in maintainable sizes.
Workaround: Manually rotate access.log to allow Squid to start.
Regards
Henrik Nordström
On Monday 17 December 2001 02.55, Andres Biscione wrote:
> Hello-
>
> We've been running Squid for a month, approximately, and we recently had
> some unprecedented problems. We don't find their cause, so maybe someone
> can help.
>
> Below is the cache.log relevant part, from the first error and on. We
> suppose the first error to be the relevant one, and the other being caused
> by it, but we're not sure.
>
> Any enlightening will be enormously appreciated.
>
> 2001/12/15 09:28:22| diskHandleWrite: FD 2: disk write error: (27) File too
> large
> 2001/12/15 09:28:22| storeDirWriteCleanLogs: Starting...
> 2001/12/15 09:28:22| WARNING: Closing open FD 17
> 2001/12/15 09:28:22| Finished. Wrote 16357 entries.
> 2001/12/15 09:28:22| Took 0.0 seconds (922196.5 entries/sec).
> FATAL: Write failure -- check your disk space and cache.log
-- MARA Systems AB Giving you basic free Squid support Priority support or Squid enhancements available on requestReceived on Sun Dec 16 2001 - 19:38:44 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 17:05:22 MST