Megaraid with brakes on?

Matthew Marlowe matt at deploylinux.net
Wed Jan 3 15:55:35 CST 2007


>> We're using RedHat 4 ES, with the ext3 FS. Our RAID controller is Dell
>> PowerEdge Expandable RAID controller 4 (rev 06) (according to lspci)
>> and we have setup a RAID level 5 with 3 SCSI disks, Seagate Barracuda
>> 10k 73gb
>> 
>> We're using the machine for serving a few million 3k files, sizing up
>> ~30gb via HTTP, using the lighttpd webserver.
>> 
>> The average io-wait for a whole day is ~18.2%, and we've some iostat
>> output that is pretty scaring (util is 98.80%).
>> 

Are all these 3K sized files accessed equally or do 10% of them account for
90% of traffic?  How much traffic is the server getting?  What is the output
of the "free" command? -- how much of your ram is dedicated to filesystem
cache?

It looks like you have 4GB ram, but my guess is that there isnt enough ram
to hold the files in cache and if the files are being accessed mostly
randomly you are going to be limited by the number of disks you have, not
the raid controller.  Each time a new file is requested, it is booting an old
one out of memory which the controller may have to go find just a few seconds
later.

The above is just my wild guess.  But, it appeared that most of your raid activity
was reads -- not writes, which would support it.



More information about the Linux-PowerEdge mailing list