![]() |
|
|
|
|
|||||||
| Forums | Portal | Register | Forum Rules | FAQ | Contribute | Members List | Arcade | Search | Today's Posts | Mark Forums Read |
| Linux RedHat, Ubuntu, SUSE, Fedora, Debian, Mandriva, Slackware, Gentoo linux, PCLinuxOS. All Linux questions here! |
|
|
||||
| Thread | Thread Starter | Forum | Replies | Last Post |
| Problem with accessing SAN disks | michael.chow | SUN Solaris | 15 | 05-22-2008 05:50 AM |
| Problem for restoring lvm on different disks | donaldfung | Linux | 0 | 08-23-2006 02:49 AM |
| Disks on AIX | piooooter | UNIX for Dummies Questions & Answers | 2 | 03-07-2006 02:20 AM |
| Solaris x86 - 2 new disks | frustrated1 | SUN Solaris | 2 | 07-07-2005 03:30 AM |
| how many disks do I have? | g-e-n-o | UNIX for Dummies Questions & Answers | 8 | 11-01-2001 12:20 PM |
|
|
Submit Tools | LinkBack | Thread Tools | Display Modes |
|
#1
|
|||
|
|||
|
problem with disks on SAN
Hi
I have a linux box attched to a SAN storage from EMC with RAID 5 .I understand that it has 3g cache howver a 20gb file creation takes too much time here are my results any ideas why time dd if=/dev/zero of=disk.img bs=1048576 count=20000 20000+0 records in 20000+0 records out 997.59s real 0.01s user 268.61s system time dd if=/dev/zero of=disk.img bs=1048576 count=20000 20000+0 records in 20000+0 records out 1093.96s real 0.04s user 269.08s system How do i go about and pin point where the problem is. regards Hrishy |
| Forum Sponsor | ||
|
|
|
#2
|
|||
|
|||
|
hmm.. Why would you consider it "too much"? I see it reasonable. The peak transfer rate to SAN (assuming you are using fiber link) is 100MB/s which is the ceiling. That leads you to 200s for 20GB data. Then, there are latencies, raid 5 parity generation, other application using SAN, etc.
So a bit slow but not unreasonable. In order to speed up, you may want to try smaller block size, say 8k or 16k. I would expect a ~100s less in real time with smaller block size. Hope it helpful, Tom |
|
#3
|
||||
|
||||
|
I agree the blocksize may be part of the issue but you may also want to check your EMC. EMCs are funny beasts, if your RAID group contains too few disks you may have a problem. We ran into this in our AIX 5.3->EMC Clariion environment.
Other things worth checking are your firmware and driver versions for your HBA. |
||||
| Google The UNIX and Linux Forums |