Niall Doherty wrote:
> I've got 7 TCP_SWAPFAIL_MISSes in my access.log file since this
> morning and 67 is yesterdays log. The only storeSwapInFileOpened
> mismatch msgs in the cache.log files are the 2 from yesterday.
There is beleived to be a race condition that may cause cache files to
be deleted, causing TCP_SWAPFAIL_MISSes.
It is also true that during a rebuild the cache index may be slightly
corrupted. Things should clear themselves when the rebuild completes.
1.2 guarantees that corrupt data is NOT sent to the client. This was
especially true in earlier 1.2beta releases if you used async-io.
There is no known issues that could cause a swapin size mismatch after
the rebuild process has completed, but there may well be a unknown bug.
Are you using async I/O?
--- Henrik Nordström Sparetime Squid HackerReceived on Sun Sep 27 1998 - 07:16:45 MDT
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:42:12 MST