Re: RAID or multiple cache_dir s?

On Thu, 31 Dec 1998, Martin Bene wrote:

]> o For a reasonably well used cache, is using multiple cache_dir entries and
]> efficient as a hardare RAID-0 box? (Yes I know there will be some
]> small performance loss, but will squid make use of all the disks
]> efficiently?)
]
]Using multiple cache dirs is much preferable to using RAIND-0: Raid 0
]doesn't have any redundancy, so you lose your complete cache if a single
]disk dies. Using multiple cache dirs, you just lose the data stored on that
]single disk. Squid spreads data evenly over all the tisks, so performance
]should be ok in normal configuration. One difference is that a single file
]gets stored to a single disk by squid, so if you've mostly got requests for
]very large files raid would be better.

If memory serves me correctly, Steward Forster believed (w/o proove) that
a "decent RAID" (I guess that means reliable HW by a well-known
manufacturer, enough NVRAM and all the goodies) would outperform the
multiple disks approach. We are talking about HW RAID 0, of course. Do
avoid SW-RAID - even though I actually saw a (admittedly special)
metadrive setup which was performing as good as the squid maintained
disks.

Losing all the data in case of a RAID0 disk failure does *not* frighten
me, we are talking about cache service after all. If one cache fails,
there should be others at the same level to take up the load. And
refilling the disks after a restart goes quickly enough.

Another thread in this discussion mentioned RAID[45]. As caches are write
intensive, and with RAID5 it was recommended to have no more than one
third writes (is that still true?), I wouldn't use RAID5 for caches. One
uses RAID5 for valuable content, but cache content can always be reclaimed
from the origin site.

Le deagh dhùrachd,
Dipl.-Ing. Jens-S. Vöckler (voeckler@rvs.uni-hannover.de)
Institute for Computer Networks and Distributed Systems
University of Hanover, Germany; +49 511 762 4726
Received on Mon Jan 04 1999 - 03:28:37 MST