Mike Diggins wrote:
> My initial thought was that a stripe set was the obvious choice for the
> squid cache. It seems that isn't the case. Why wouldn't Squid benefit from
> the stripe set?
Because Squid performs small random I/O aldready distributed over the
cache_dir lines you have. It does not need the help of a stripe manager
to distribute the I/O load on more than one disk.
Striped disks are best at high bandwidth sequential read/writes. Squid
performs none of these, nor does it require high disk I/O bandwidth.
What Squid requires is good seek times which you get by having multiple
drives being able to service different requests in parallell.
> Also, if I have, say, two cache dirs spread over two physical disks and
> one drive fails, will squid continue to run with one cache dir out of
> commission?
Yes, but you might have to remove the crashed disks cache_dir line from
squid.conf before Squid is fully happy.
As robert said there is also an effort on squid.sourceforge.net to
improve on this to allow "hot" addition/removal of cache directories,
but the status of this work is unknown.
-- Henrik Nordstrom Squid HackerReceived on Sun Apr 29 2001 - 04:12:02 MDT
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:59:39 MST