Re: Squid 2.0 PATCH 2 and cache dir's filling up.

On Tue, 27 Oct 1998, Henrik Nordstrom wrote:

> John Sloan wrote:
> > it threw away most of the entries in the swaplog file without
> > succeeding in deleteing most of the contents of the disk.

> Sounds like it failed to write the index file.

> Repairing a lost index file is easy. Simply shut down Squid, delete the
> index file and start Squid again.

When squid does this, how does it handle the cache churn while rebuilding
the index.

Or does it just not cache or purge any objects during the building of the
swap.state?

What I was worried about when I was faced with a similar situation to
John, was that during the day(s) it would take to rebuild the swap.state
for our 4 x 6Gb drives, is there going to be a discrepency with the
swap.state it comes up with, compared to the real snapshot on the
drive(s) due to churn of objects from clients accessing and using the
squid server.

I opted for wiping all the drives and starting again (my swap.state's were
around 12k, when normally they are ~20Mb [filesystem thought it had run
out of room and the swap.state's were corrupted]. fsck fixed the drive
confusion).

-----
Roger Yerramsetti : ,-------. :
rogery@wantree.com.au : : / / | a n t r e e I n t e r n e t :
Network Administrator : | (_/_/`-'----------------------------- :
Ph (08) 9221 8899 ::::: `-------' http://www.wantree.com.au ::::
Received on Tue Oct 27 1998 - 07:50:01 MST