On Thu, 25 May 2006, Steve Shockley wrote:
Jeff Ross wrote:
Backing up root filesystem:
copying /dev/rsd0a to /dev/rsd0n
One spindle, disk thrashes.
Backing up root filesystem:
copying /dev/rwd0a to /dev/rwd1m
Two spindles, disk does not thrash.
I wondered about that, but here is another, perhaps better, example:
Backing up root filesystem:
copying /dev/rwd0a to /dev/rwd0l
65515+1 records in
65515+1 records out
536699392 bytes transferred in 126.647 secs (4237751 bytes/sec)
** /dev/rwd0l
** Last Mounted on /
** Phase 1 - Check Blocks and Sizes
** Phase 2 - Check Pathnames
** Phase 3 - Check Connectivity
** Phase 4 - Check Reference Counts
UNREF FILE I=129 OWNER=root MODE=140666
SIZE=0 MTIME=May 7 16:38 2006
CLEAR? yes
Same spindle, but 8x as fast as the 10k SCSI disks.
But, that must be part of it at least:
/dev/sd2b 4.9G 1.3G 3.3G 28% /applications
/dev/sd1d 7.0G 2.5G 4.1G 38% /shared
[EMAIL PROTECTED]:/applications $ dd of=empty.txt if=/dev/zero bs=1024
count=100000
100000+0 records in
100000+0 records out
102400000 bytes transferred in 2.408 secs (42523822 bytes/sec)
[EMAIL PROTECTED]:/applications $ dd if=empty.txt of=/tmp/empty.txt bs=1024
100000+0 records in
100000+0 records out
102400000 bytes transferred in 3.156 secs (32444839 bytes/sec)
[EMAIL PROTECTED]:/applications $ dd if=empty.txt of=empty2.txt bs=1024
100000+0 records in
100000+0 records out
102400000 bytes transferred in 27.184 secs (3766863 bytes/sec)
I "inherited" this hardware and gleefully replaced NT4 to run
samba and postgresql for the local animal shelter. Would replacing the
controller help? Not that they've got the money for that ;-)
Jeff,
(who's having a sudden panic attack about the LSI MegaRAID and 7 Hitachi
10K drives freshly arrived to stuff into the server at work...)
