Hi>Looks like my RAID5 arrays have crap performance - esp. the master array.
OK here are the stats from my RAID5 setup with 4 X Seagate 160Gb Drives (all masters):Version 1.03 ------Sequential Output------ --Sequential Input- --Random-
-Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP
gryffindor 1G 23157 83 43898 24 14863 8 17742 58 32670 10 369.4 2
------Sequential Create------ --------Random Create--------
-Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
files /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
16 1598 98 +++++ +++ +++++ +++ 1640 97 +++++ +++ 4086 97
So I see the
traditional high CPU on writes but reads are fast with little overhead. Your
figures do look worse than mine in places (mostly in the small file tests,
although you seem to be using much less CPU time).
I'm not worried
about the high write times as the volumes I have on the array are mainly
read. I host my music, digital picture and movie files
there.
For myth buffers and
my db I have an XFS filesystem on a standalone disk, and my OS is on a RAID 1
mirror on another 2 disks.
Just curious if you
got any output in dmesg or /var/log/messages whilst running. I didn't see
anything but I'm guessing that if there were bus timing errors that'd be where
they'd be. Did you (dare) try running this on the master and slave arrays at the
same time ? I'm guessing it would grind to a halt.
Regards
Paul
| ||||||||||||||
| ||||||||||||||
_______________________________________________ mythtv-users mailing list [email protected] http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-users
