I have a server that is running slowly and I think it's because of my stupidity - it's running CentOS 3.8
I have software RAID level 1 and the mirrored drives are the master & slave on the primary controller and I'm thinking that is my mistake.
Is there a simple method for changing the configuration of md before I switch the slave drive to be the slave on the secondary controller?
Craig
Craig White wrote:
I have a server that is running slowly and I think it's because of my stupidity - it's running CentOS 3.8
I have software RAID level 1 and the mirrored drives are the master & slave on the primary controller and I'm thinking that is my mistake.
Is there a simple method for changing the configuration of md before I switch the slave drive to be the slave on the secondary controller?
I think they will pair up automatically if you shut down cleanly and move the connector. If not, just 'mdadm --add' it back after it comes up with a missing member. If you have to add it back, the drives will have to resync so there is a small risk of failure during the process.
On Fri, 2007-01-12 at 22:31 -0600, Les Mikesell wrote:
Craig White wrote:
I have a server that is running slowly and I think it's because of my stupidity - it's running CentOS 3.8
I have software RAID level 1 and the mirrored drives are the master & slave on the primary controller and I'm thinking that is my mistake.
Is there a simple method for changing the configuration of md before I switch the slave drive to be the slave on the secondary controller?
I think they will pair up automatically if you shut down cleanly and move the connector. If not, just 'mdadm --add' it back after it comes up with a missing member. If you have to add it back, the drives will have to resync so there is a small risk of failure during the process.
---- great thanks - am I right in thinking that having both mirrored drives on the same controller master/slave is likely a cause of slowness?
i.e.
# mdadm --query --detail /dev/md1 /dev/md1: Version : 00.90.00 Creation Time : Tue Feb 1 04:10:32 2005 Raid Level : raid1 Array Size : 2096384 (2047.25 MiB 2146.70 MB) Device Size : 2096384 (2047.25 MiB 2146.70 MB) Raid Devices : 2 Total Devices : 2 Preferred Minor : 1 Persistence : Superblock is persistent
Update Time : Sat Jan 13 13:53:23 2007 State : dirty, no-errors Active Devices : 2 Working Devices : 2 Failed Devices : 0 Spare Devices : 0
Number Major Minor RaidDevice State 0 3 6 0 active sync /dev/hda6 1 3 70 1 active sync /dev/hdb6 UUID : ff9049f4:29e82010:e70acb45:690a1ad5 Events : 0.46
Craig
Craig White wrote:
I have software RAID level 1 and the mirrored drives are the master & slave on the primary controller and I'm thinking that is my mistake.
Is there a simple method for changing the configuration of md before I switch the slave drive to be the slave on the secondary controller?
I think they will pair up automatically if you shut down cleanly and move the connector. If not, just 'mdadm --add' it back after it comes up with a missing member. If you have to add it back, the drives will have to resync so there is a small risk of failure during the process.
great thanks - am I right in thinking that having both mirrored drives on the same controller master/slave is likely a cause of slowness?
It will make some difference on writes, but reads should only be done on one of the drives anyway. I doubt if it makes a huge difference unless you have an application writing all the time.
On Fri, 2007-01-12 at 23:16 -0600, Les Mikesell wrote:
Craig White wrote:
I have software RAID level 1 and the mirrored drives are the master & slave on the primary controller and I'm thinking that is my mistake.
Is there a simple method for changing the configuration of md before I switch the slave drive to be the slave on the secondary controller?
I think they will pair up automatically if you shut down cleanly and move the connector. If not, just 'mdadm --add' it back after it comes up with a missing member. If you have to add it back, the drives will have to resync so there is a small risk of failure during the process.
great thanks - am I right in thinking that having both mirrored drives on the same controller master/slave is likely a cause of slowness?
It will make some difference on writes, but reads should only be done on one of the drives anyway. I doubt if it makes a huge difference unless you have an application writing all the time.
---- here's my problem - system was ok and mother board died. Purchased a new motherboard and this is what I get...
# hdparm -t /dev/md1
/dev/md1: Timing buffered disk reads: 4 MB in 3.68 seconds = 1.09 MB/sec
which is dreadfully slow
Thanks
Craig
-----Original Message----- From: centos-bounces@centos.org [mailto:centos-bounces@centos.org] On Behalf Of Craig White Sent: Saturday, January 13, 2007 12:40 AM To: CentOS mailing list Subject: Re: [CentOS] slow system
On Fri, 2007-01-12 at 23:16 -0600, Les Mikesell wrote:
Craig White wrote:
I have software RAID level 1 and the mirrored drives
are the master &
slave on the primary controller and I'm thinking that
is my mistake.
Is there a simple method for changing the configuration
of md before I
switch the slave drive to be the slave on the secondary
controller?
I think they will pair up automatically if you shut down cleanly and move the connector. If not, just 'mdadm --add' it back after it comes up with a missing member. If you have to add it back, the drives will have to resync so there is a small risk of failure during the process.
great thanks - am I right in thinking that having both
mirrored drives
on the same controller master/slave is likely a cause of slowness?
It will make some difference on writes, but reads should only be done on one of the drives anyway. I doubt if it makes a huge difference unless you have an application writing all the time.
here's my problem - system was ok and mother board died. Purchased a new motherboard and this is what I get...
# hdparm -t /dev/md1
/dev/md1: Timing buffered disk reads: 4 MB in 3.68 seconds = 1.09 MB/sec
which is dreadfully slow
Something is definitely not right here. I have an old Dell Dimension (1GHz, 512MB) at home here that I have FC5 on, the drives are on the first channel master/slave mirrored, cheapo 100GB WDs. Here is what I get:
[root@gandalf linux]# hdparm -t /dev/md1
/dev/md1: Timing buffered disk reads: 72 MB in 3.00 seconds = 23.96 MB/sec
Now that is what I call slow, what you are getting is unusable.
I'd put then on separate channels, but I have an LS-120 drive in PIO, and a CDRW running at like DMA2, and it would be a lot worse to share a channel with one of those.
-Ross
______________________________________________________________________ This e-mail, and any attachments thereto, is intended only for use by the addressee(s) named herein and may contain legally privileged and/or confidential information. If you are not the intended recipient of this e-mail, you are hereby notified that any dissemination, distribution or copying of this e-mail, and any attachments thereto, is strictly prohibited. If you have received this e-mail in error, please immediately notify the sender and permanently delete the original and any copy or printout thereof.
Craig White wrote:
great thanks - am I right in thinking that having both mirrored drives on the same controller master/slave is likely a cause of slowness?
It will make some difference on writes, but reads should only be done on one of the drives anyway. I doubt if it makes a huge difference unless you have an application writing all the time.
here's my problem - system was ok and mother board died. Purchased a new motherboard and this is what I get...
# hdparm -t /dev/md1
/dev/md1: Timing buffered disk reads: 4 MB in 3.68 seconds = 1.09 MB/sec
which is dreadfully slow
That's worse than slow. Something is broken. Do the drives perform better if you test /dev/hda or /dev/hdb? What does dmesg say about them - does it mention UDMA(100) or better? Are you using a new-style 80-wire cable? It might be worth buying an IDE controller card if the on-board one won't do better than that.
On Sat, 2007-01-13 at 00:08 -0600, Les Mikesell wrote:
Craig White wrote:
great thanks - am I right in thinking that having both mirrored drives on the same controller master/slave is likely a cause of slowness?
It will make some difference on writes, but reads should only be done on one of the drives anyway. I doubt if it makes a huge difference unless you have an application writing all the time.
here's my problem - system was ok and mother board died. Purchased a new motherboard and this is what I get...
# hdparm -t /dev/md1
/dev/md1: Timing buffered disk reads: 4 MB in 3.68 seconds = 1.09 MB/sec
which is dreadfully slow
That's worse than slow. Something is broken. Do the drives perform better if you test /dev/hda or /dev/hdb?
---- no # hdparm -t /dev/hda6
/dev/hda6: Timing buffered disk reads: 4 MB in 3.63 seconds = 1.10 MB/sec [root@spot root]# hdparm -t /dev/hdb6
/dev/hdb6: Timing buffered disk reads: 4 MB in 3.60 seconds = 1.11 MB/sec ----
What does dmesg say about them - does it mention UDMA(100) or better? Are you using a new-style 80-wire cable? It might be worth buying an IDE controller card if the on-board one won't do better than that.
---- dmesg states nothing about ATA 100...
Uniform Multi-Platform E-IDE driver Revision: 7.00beta4-2.4 ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx hda: WDC WD800JB-00JJA0, ATA DISK drive hdb: HDS728080PLAT20, ATA DISK drive
I'm going over there tomorrow and I will surely bring a new 80-wire cable.
Thanks
Craig
On Sat, 2007-01-13 at 16:30 -0500, Stephen Harris wrote:
On Sat, Jan 13, 2007 at 02:18:12PM -0700, Craig White wrote:
/dev/hda6: Timing buffered disk reads: 4 MB in 3.63 seconds = 1.10 MB/sec
What does "hdparm -i /dev/hda" say?
---- # hdparm -i /dev/hda
/dev/hda:
Model=WDC WD800JB-00JJA0, FwRev=05.01C05, SerialNo=WD-WCAM92819556 Config={ HardSect NotMFM HdSw>15uSec SpinMotCtl Fixed DTR>5Mbs FmtGapReq } RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=66 BuffType=unknown, BuffSize=8192kB, MaxMultSect=16, MultSect=16 CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=156301488 IORDY=on/off, tPIO={min:120,w/IORDY:120}, tDMA={min:120,rec:120} PIO modes: pio0 pio3 pio4 DMA modes: mdma0 mdma1 mdma2 UDMA modes: udma0 udma1 udma2 udma3 udma4 *udma5 AdvancedPM=no WriteCache=enabled Drive conforms to: device does not report version:
* signifies the current active mode
Craig
Craig White wrote:
On Sat, 2007-01-13 at 16:30 -0500, Stephen Harris wrote:
On Sat, Jan 13, 2007 at 02:18:12PM -0700, Craig White wrote:
/dev/hda6: Timing buffered disk reads: 4 MB in 3.63 seconds = 1.10 MB/sec
What does "hdparm -i /dev/hda" say?
# hdparm -i /dev/hda
/dev/hda:
Model=WDC WD800JB-00JJA0, FwRev=05.01C05, SerialNo=WD-WCAM92819556 Config={ HardSect NotMFM HdSw>15uSec SpinMotCtl Fixed DTR>5Mbs FmtGapReq } RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=66 BuffType=unknown, BuffSize=8192kB, MaxMultSect=16, MultSect=16 CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=156301488 IORDY=on/off, tPIO={min:120,w/IORDY:120}, tDMA={min:120,rec:120} PIO modes: pio0 pio3 pio4 DMA modes: mdma0 mdma1 mdma2 UDMA modes: udma0 udma1 udma2 udma3 udma4 *udma5 AdvancedPM=no WriteCache=enabled Drive conforms to: device does not report version:
- signifies the current active mode
phew, something is *way* wrong. this is from a p3-450Mhz RHL 6.3 system I have at home, ancient Intel 440BX motherboard, and an ancient IBM Deskstar 25GB 5400rpm drive.
# hdparm -t /dev/hda
/dev/hda: Timing buffered disk reads: 64 MB in 4.44 seconds = 14.41 MB/sec
# hdparm -i /dev/hda
/dev/hda:
Model=IBM-DJNA-352500, FwRev=J51OA30K, SerialNo=GW0GWFD6488 Config={ HardSect NotMFM HdSw>15uSec Fixed DTR>10Mbs } RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=34 BuffType=3(DualPortCache), BuffSize=1966kB, MaxMultSect=16, MultSect=16 DblWordIO=no, OldPIO=2, DMA=yes, OldDMA=2 CurCHS=16383/16/63, CurSects=-66060037, LBA=yes, LBAsects=49981680 tDMA={min:120,rec:120}, DMA modes: mword0 mword1 mword2 IORDY=on/off, tPIO={min:240,w/IORDY:120}, PIO modes: mode3 mode4 UDMA modes: mode0 mode1 *mode2 mode3 mode4
# uname -a Linux hogranch.com 2.2.24-6.2.3 #1 Fri Mar 14 08:41:15 EST 2003 i686 unknown
here's a p3 866Mhz running CentOS 4 on a crappy i810e (old dell optiplex) w/ an old Seagate 10GB 7200rpm
# uname -a Linux svfis-linux1 2.6.9-42.0.3.EL #1 Fri Oct 6 05:59:54 CDT 2006 i686 i686 i386 GNU/Linux
# hdparm -t /dev/hda
/dev/hda: Timing buffered disk reads: 82 MB in 3.05 seconds = 26.92 MB/sec
# hdparm -i /dev/hda
/dev/hda:
Model=ST310210A, FwRev=3.17, SerialNo=3CM0T2X9 Config={ HardSect NotMFM HdSw>15uSec Fixed DTR>10Mbs RotSpdTol>.5% } RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=0 BuffType=unknown, BuffSize=2048kB, MaxMultSect=16, MultSect=16 CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=19925880 IORDY=on/off, tPIO={min:240,w/IORDY:120}, tDMA={min:120,rec:120} PIO modes: pio0 pio1 pio2 pio3 pio4 DMA modes: mdma0 mdma1 mdma2 UDMA modes: udma0 udma1 udma2 udma3 *udma4 AdvancedPM=no WriteCache=enabled Drive conforms to: device does not report version: 1 2 3 4
* signifies the current active mode
and, a IBM BladeCenter Xeon 2.8Ghz w/ a 2.5" SCSI 36GB...
# hdparm -t /dev/sda /dev/sda: Timing buffered disk reads: 160 MB in 3.20 seconds = 50.01 MB/sec
On 14/01/07, Craig White craigwhite@azapple.com wrote:
here's my problem - system was ok and mother board died. Purchased a
new
motherboard and this is what I get...
# hdparm -t /dev/md1
/dev/md1: Timing buffered disk reads: 4 MB in 3.68 seconds = 1.09 MB/sec
which is dreadfully slow
Any decent system with PATA drives should show values in range 30~50 MB/sec. It would be more useful to get output from "hdparm -tT /dev/(yourdrive)"
Shots in the dark. -- Does this system have SiL drive controller chip? I ran into similar out put problem with a MSI MoBo having that chipset two years back although with SATA drives. -- Replace drive cables. I have seen degraded output with some old cables but never so much.
HTH
On Sun, 2007-01-14 at 08:16 +0530, Sudev Barar wrote:
On 14/01/07, Craig White craigwhite@azapple.com wrote: > > here's my problem - system was ok and mother board died. Purchased a new > > motherboard and this is what I get... > > > > # hdparm -t /dev/md1 > > > > /dev/md1: > > Timing buffered disk reads: 4 MB in 3.68 seconds = 1.09 MB/sec > > > > which is dreadfully slow
Any decent system with PATA drives should show values in range 30~50 MB/sec. It would be more useful to get output from "hdparm -tT /dev/(yourdrive)"
Shots in the dark. -- Does this system have SiL drive controller chip? I ran into similar out put problem with a MSI MoBo having that chipset two years back although with SATA drives.
-- Replace drive cables. I have seen degraded output with some old cables but never so much.
---- it was an Intel M/B and putting in a separate ATA controller fixed it (it wasn't the drive cable or the drives) - this ended up moving the drives to hde/hdf respectively and I ended up having to rebuild the faulty mirrored drives - easy enough, I didn't ever have to do that before but it wasn't difficult.
Thanks Les/Sudev/all
Craig
ok which mobo did you get? You could have an issue with the mobo's chipset support. Which version of Centos?
Craig White wrote:
On Fri, 2007-01-12 at 23:16 -0600, Les Mikesell wrote:
Craig White wrote:
I have software RAID level 1 and the mirrored drives are the master & slave on the primary controller and I'm thinking that is my mistake.
Is there a simple method for changing the configuration of md before I switch the slave drive to be the slave on the secondary controller?
I think they will pair up automatically if you shut down cleanly and move the connector. If not, just 'mdadm --add' it back after it comes up with a missing member. If you have to add it back, the drives will have to resync so there is a small risk of failure during the process.
great thanks - am I right in thinking that having both mirrored drives on the same controller master/slave is likely a cause of slowness?
It will make some difference on writes, but reads should only be done on one of the drives anyway. I doubt if it makes a huge difference unless you have an application writing all the time.
here's my problem - system was ok and mother board died. Purchased a new motherboard and this is what I get...
# hdparm -t /dev/md1
/dev/md1: Timing buffered disk reads: 4 MB in 3.68 seconds = 1.09 MB/sec
which is dreadfully slow
Thanks
Craig
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
great thanks - am I right in thinking that having both mirrored drives on the same controller master/slave is likely a cause of slowness?
that would depend if the channel or the drive is the bottleneck. if the channel is fast enough to service both devices running at speed, then, no problem.
-
Craig White -
John R Pierce -
Les Mikesell -
Ross S. W. Walker -
Stephen Harris -
Sudev Barar -
William Warren