>>> Chris Tilbury wrote
> You're quite right in that if you are using each disk as a separate entity,
> then if you lose one then you can still preserve the remainder of the cache,
> though. There would certainly be no point to mirroring individual disks in
> this scenario.
And really, who can justify mirroring cache disks? Even if you're paying
for every MB of data, you've got to be using some pretty flaky disks
before the bandwidth costs outweigh the costs of the disks and the extra
controllers (since you're not going to mirror with both disks on the
same controller - that would be stupid)
> cause a performance degradation. Using a transaction log (or NVRAM if you
> are very rich) _will_ help this. Sure, you can often turn many of these off
> by disabling synchronous writes of metadata, but that can have adverse side
> effects too.
One of the problems with the newer pieces of Sun hardware (eg the U250 or
U450) versus the older U2s is that you can't get Sun to support Prestoserve
on the newer ones. There's undoubtedly PCI prestoserve out there, but no
support for Solaris. :(
> True, but it does matter if the machine can't properly restart because it
> cannot clean the cache filesystem(s). This is the risk you increase vastly
> if you disable synchronous metadata updates to get the performance increase.
Nah - if it's really a problem, change the boot up scripts to trash the
disk if it won't fsck. Or just run "fsck -y" in the boot up (danger: not
recommended on any disk that you care about :)
If your caches are crashing at all, you have a problem, anyway.
> I think the latter certainly needs to have at least token attention paid to
> it. Whilst issues related to crash recovery might not be so important on
> simple cache "appliance" type systems, on a cache running on a general
> purpose operating system with many more "moving parts", it's of far more
> relevance.
It's being addressed - part of the goal is to make the thing as simple as
possible, which will help the crash recovery.
> Perhaps some of the work done with INN2 on CNFS could be of value here, as
> news seems to at least superficially similar in its use of filesystems for
> object storage (large numbers of directories, with clusters of generally
> small files).
Nope, it's not really - the expiry patterns of the filesystems are totally
different.
Anthony
Received on Tue Sep 15 1998 - 01:36:47 MDT
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:42:01 MST