Recommended NAS?

[KJake]

Originally Posted by mikesm

That smells about right to me. It would be nice if it went faster, but since the volume is online and really not all that impacted performance wise, it's not that big a deal to me...

Yeah, a dd would be much faster, but it doesn't take care to lock out things so the data doesn't change while it is being moved.

58.5% now, 18 hours in.

[KJake]

Originally Posted by KJake

Yeah, a dd would be much faster, but it doesn't take care to lock out things so the data doesn't change while it is being moved.

58.5% now, 18 hours in.

Operation completed successfully!

I used EVMS to remove the old /dev/md0 and all the disk segments and then grew the new array by 465GB. Flawless. I love this.

[KJake]

I took my 6x500 disks out today and I noticed some things.

First, when I recompiled my kernel, I didn't include the bd-claim patch like I had in my other kernel. This caused a little problem because I had some mixed mount references in my fstab and the udev device-mapper and evms started fighting. So I switched everying in fstab to UUID mounting and I think that fixed that problem, but it looks like my SWAP won't mount automatically. I was only able to turn it on via EVMS. How are others doing this? Do you have EVMS managing your root(/) and swap?

/etc/fstab:
UUID=f3d99fa0-a30a-4083-918a-e30558361d0e       none    swap    sw      0 0

Second, this might be more for mikesm since I know he has one of these. I have one of those addonics 5-in-1 enclosures connected to a addonics PMP. When the system starts up, it issues hard resets to all the PMP ports (including my other PMP). I think this is to get the ports configured for the fastest speed possible. Well, the 5-in-1 connected to the PMP doesn't seem to like it and I was wondering if you see the same thing, or if this is an indication of something wrong? I'm not worried because it does eventually get there, it just slows down the boot time a bit while it waits for it to respond properly.

dmesg output:
[   97.707760] ata2: SATA link up 3.0 Gbps (SStatus 123 SControl 0)
[   97.707936] ata2.15: Port Multiplier 1.1, 0x1095:0x3726 r23, 6 ports, feat 0x1/0x9
[   97.708153] ata2.00: hard resetting link
[   98.027519] ata2.00: SATA link down (SStatus 0 SControl 10)
[   98.027542] ata2.01: hard resetting link
[   98.475216] ata2.01: softreset failed (SRST command error)
[   98.475254] ata2.01: reset failed (errno=-5), retrying in 10 secs
[  108.025205] ata2.01: hard resetting link
[  108.025228] ata2.01: failed to read SCR 2 (Emask=0x40)
[  108.025230] ata2.01: failed to read SCR 2 (Emask=0x40)
[  108.025231] ata2.01: COMRESET failed (errno=-5)
[  108.025258] ata2.01: failed to read SCR 0 (Emask=0x40)
[  108.025259] ata2.01: reset failed, giving up
[  108.025286] ata2.15: hard resetting link
[  108.025288] ata2: controller in dubious state, performing PORT_RST
[  110.256753] ata2.15: SATA link up 3.0 Gbps (SStatus 123 SControl 0)
[  110.256948] ata2.00: hard resetting link
[  110.576880] ata2.00: SATA link down (SStatus 0 SControl 10)
[  110.576903] ata2.01: hard resetting link
[  111.024575] ata2.01: softreset failed (SRST command error)
[  111.024610] ata2.01: reset failed (errno=-5), retrying in 10 secs
[  120.574564] ata2.01: hard resetting link
[  120.574587] ata2.01: failed to read SCR 2 (Emask=0x40)
[  120.574588] ata2.01: failed to read SCR 2 (Emask=0x40)
[  120.574589] ata2.01: COMRESET failed (errno=-5)
[  120.574616] ata2.01: failed to read SCR 0 (Emask=0x40)
[  120.574617] ata2.01: reset failed, giving up
[  120.574645] ata2.15: hard resetting link
[  120.574646] ata2: controller in dubious state, performing PORT_RST
[  122.806112] ata2.15: SATA link up 3.0 Gbps (SStatus 123 SControl 0)
[  122.806305] ata2.00: hard resetting link
[  123.126240] ata2.00: SATA link down (SStatus 0 SControl 10)
[  123.126263] ata2.01: hard resetting link
[  123.478110] ata2.01: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[  123.478133] ata2.02: hard resetting link
[  123.925859] ata2.02: softreset failed (SRST command error)
[  123.925895] ata2.02: reset failed (errno=-5), retrying in 10 secs
[  133.475849] ata2.02: hard resetting link
[  133.587830] ata2.02: hardreset failed (port not ready)
[  133.587858] ata2.02: reset failed (errno=-5), retrying in 10 secs
[  143.473744] ata2.02: hard resetting link
[  143.473767] ata2.02: failed to read SCR 2 (Emask=0x40)
[  143.473769] ata2.02: failed to read SCR 2 (Emask=0x40)
[  143.473770] ata2.02: COMRESET failed (errno=-5)
[  143.473796] ata2.02: failed to read SCR 0 (Emask=0x40)
[  143.473797] ata2.02: reset failed, giving up
[  143.473825] ata2.15: hard resetting link
[  143.473826] ata2: controller in dubious state, performing PORT_RST
[  145.705292] ata2.15: SATA link up 3.0 Gbps (SStatus 123 SControl 0)
[  145.705485] ata2.00: hard resetting link
[  146.025418] ata2.00: SATA link down (SStatus 0 SControl 10)
[  146.025441] ata2.01: hard resetting link
[  146.377289] ata2.01: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[  146.377311] ata2.02: hard resetting link
[  146.729219] ata2.02: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[  146.729242] ata2.03: hard resetting link
[  147.176965] ata2.03: softreset failed (SRST command error)
[  147.177001] ata2.03: reset failed (errno=-5), retrying in 10 secs
[  156.726955] ata2.03: hard resetting link
[  156.838932] ata2.03: hardreset failed (port not ready)
[  156.838963] ata2.03: reset failed (errno=-5), retrying in 10 secs
[  166.724851] ata2.03: hard resetting link
[  166.724873] ata2.03: failed to read SCR 2 (Emask=0x40)
[  166.724875] ata2.03: failed to read SCR 2 (Emask=0x40)
[  166.724876] ata2.03: COMRESET failed (errno=-5)
[  166.724905] ata2.03: failed to read SCR 0 (Emask=0x40)
[  166.724907] ata2.03: reset failed, giving up
[  166.724937] ata2.15: hard resetting link
[  166.724938] ata2: controller in dubious state, performing PORT_RST
[  168.956398] ata2.15: SATA link up 3.0 Gbps (SStatus 123 SControl 0)
[  168.956593] ata2.00: hard resetting link
[  169.276524] ata2.00: SATA link down (SStatus 0 SControl 10)
[  169.276547] ata2.01: hard resetting link
[  169.628398] ata2.01: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[  169.628421] ata2.02: hard resetting link
[  169.980322] ata2.02: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[  169.980345] ata2.03: hard resetting link
[  170.332249] ata2.03: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[  170.332272] ata2.04: hard resetting link
[  170.779998] ata2.04: softreset failed (SRST command error)
[  170.780036] ata2.04: reset failed (errno=-5), retrying in 10 secs
[  180.329989] ata2.04: hard resetting link
[  180.330012] ata2.04: failed to read SCR 2 (Emask=0x40)
[  180.330013] ata2.04: failed to read SCR 2 (Emask=0x40)
[  180.330014] ata2.04: COMRESET failed (errno=-5)
[  180.330044] ata2.04: failed to read SCR 0 (Emask=0x40)
[  180.330045] ata2.04: reset failed, giving up
[  180.330075] ata2.15: hard resetting link
[  180.330076] ata2: controller in dubious state, performing PORT_RST
[  182.561534] ata2.15: SATA link up 3.0 Gbps (SStatus 123 SControl 0)
[  182.561729] ata2.00: hard resetting link
[  182.881661] ata2.00: SATA link down (SStatus 0 SControl 10)
[  182.881684] ata2.01: hard resetting link
[  183.233534] ata2.01: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[  183.233557] ata2.02: hard resetting link
[  183.585460] ata2.02: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[  183.585483] ata2.03: hard resetting link
[  183.937386] ata2.03: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[  183.937409] ata2.04: hard resetting link
[  184.289312] ata2.04: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[  184.289314] ata2.05: hard resetting link
[  184.609299] ata2.05: SATA link up 1.5 Gbps (SStatus 113 SControl 320)
[  184.627251] ata2.01: ATA-7: Hitachi HDS721010KLA330, GKAOA70M, max UDMA/133
[  184.627253] ata2.01: 1953525168 sectors, multi 0: LBA48 NCQ (depth 31/32)
[  184.628139] ata2.01: configured for UDMA/100
[  184.642963] ata2.02: ATA-7: Hitachi HDS721010KLA330, GKAOA70M, max UDMA/133
[  184.642965] ata2.02: 1953525168 sectors, multi 0: LBA48 NCQ (depth 31/32)
[  184.643850] ata2.02: configured for UDMA/100
[  184.658521] ata2.03: ATA-7: Hitachi HDS721010KLA330, GKAOA70F, max UDMA/133
[  184.658523] ata2.03: 1953525168 sectors, multi 0: LBA48 NCQ (depth 31/32)
[  184.659409] ata2.03: configured for UDMA/100
[  184.660972] ata2.04: ATA-8: ST31000340AS, SD15, max UDMA/133
[  184.660974] ata2.04: 1953525168 sectors, multi 0: LBA48 NCQ (depth 31/32)
[  184.662933] ata2.04: configured for UDMA/100
[  184.663127] ata2.00: exception Emask 0x10 SAct 0x0 SErr 0x0 action 0x1 t4
[  184.663158] ata2.00: irq_stat 0x00060002, failed to transmit command FIS
[  184.663254] ata2: EH complete

[mikesm]

Originally Posted by KJake

I took my 6x500 disks out today and I noticed some things.

First, when I recompiled my kernel, I didn't include the bd-claim patch like I had in my other kernel. This caused a little problem because I had some mixed mount references in my fstab and the udev device-mapper and evms started fighting. So I switched everying in fstab to UUID mounting and I think that fixed that problem, but it looks like my SWAP won't mount automatically. I was only able to turn it on via EVMS. How are others doing this? Do you have EVMS managing your root(/) and swap?

/etc/fstab:
UUID=f3d99fa0-a30a-4083-918a-e30558361d0e       none    swap    sw      0 0

Second, this might be more for mikesm since I know he has one of these. I have one of those addonics 5-in-1 enclosures connected to a addonics PMP. When the system starts up, it issues hard resets to all the PMP ports (including my other PMP). I think this is to get the ports configured for the fastest speed possible. Well, the 5-in-1 connected to the PMP doesn't seem to like it and I was wondering if you see the same thing, or if this is an indication of something wrong? I'm not worried because it does eventually get there, it just slows down the boot time a bit while it waits for it to respond properly.

I can't address the kernel patch as the patched kernel in opensuse 10.3 has what I need in it, but then I use EVMS as well.

On the start-up process, it's not just trying to negotiate the highest speed, but if you listen to the drives during start up, I think you'll notice them spinning up one by one during this process. The reason you get the reset, try again, it works, try the next drive, reset, etc... is because the drive hasn't been spun up. It then spins it up, tries again, works, then goes on to the next drive.

The driver does this by default, but there may be a compile time option to try and spin everything up at once (I actually don't think this is wise, esp. with a lot of drives)... It seems to do this even on non-5in1 racks, though I can't be sure of it...

[KJake]

Originally Posted by mikesm

On the start-up process, it's not just trying to negotiate the highest speed, but if you listen to the drives during start up, I think you'll notice them spinning up one by one during this process. The reason you get the reset, try again, it works, try the next drive, reset, etc... is because the drive hasn't been spun up. It then spins it up, tries again, works, then goes on to the next drive.

Yup, I can confirm this. The times that I noticed this had been from cold boot, I just rebooted the system and didn't see all the "errors" this time. Thanks!

[mikesm]

Originally Posted by KJake

Yup, I can confirm this. The times that I noticed this had been from cold boot, I just rebooted the system and didn't see all the "errors" this time. Thanks!

Staggered spinup is a very good thing in NAS's in general. It adds about 3 mins to my 13 drive configuration, but lets me run everything with a 480W PSU. If all 13 drives spun up at once, I'd need a huge PSU to deal with the surges on the +12V line.

[bcjenkins]

My current configuration is 6 500GB SataII on ICH9R running 4 in Raid 5 XFS and 2 in Raid 0 (TV) XFS. I am wondering if I am doing myself a disservice by not running all in raid 5, aside from the redundancy? I am trying to solve an issue with multiple hdstreams being captured simul. which causes blips.

md1 is the raid0

root@sagetv-server:/home/bcjenkins# hdparm -tT /dev/md1

/dev/md1:
Timing cached reads: 7882 MB in 2.00 seconds = 3944.53 MB/sec
Timing buffered disk reads: 542 MB in 3.05 seconds = 177.93 MB/sec

root@sagetv-server:/home/bcjenkins# hdparm -tT /dev/mapper/video_vg-video_lv

/dev/mapper/video_vg-video_lv:
Timing cached reads: 7598 MB in 2.00 seconds = 3802.10 MB/sec
Timing buffered disk reads: 360 MB in 3.01 seconds = 119.44 MB/sec
root@sagetv-server:/home/bcjenkins#

Thanks in advance

B

[mikesm]

Originally Posted by bcjenkins

My current configuration is 6 500GB SataII on ICH9R running 4 in Raid 5 XFS and 2 in Raid 0 (TV) XFS. I am wondering if I am doing myself a disservice by not running all in raid 5, aside from the redundancy? I am trying to solve an issue with multiple hdstreams being captured simul. which causes blips.

md1 is the raid0



Thanks in advance

B

I think you'd be better off running them all in RAID5, but even so 6 of them should be plenty fast. That should not be the source of any recording problems. Are you running Sage on the same hardware, or is it going over the network?

Did you try modifying the diskopt.sh script for your setup and running that? It helped quite a bit with filesystem performance, as the disk params are not tuned for raid by default.

Also, might you be using segate or maxtor disks? If so, you need to make sure the disks don't have the dreaded AAK firmware in them which kills raid performance, and that they are all the same make and model (at least pretty much). Heterogeneous disks in RAID is a nono, though turning off NCQ as the diskopt.sh script does help.

In any case, while those disk I/O numbers while slower than they could be, are still plenty fine for a Sage system to record and playback on. If you are running Sage over a network to this system, I would look at your network setup and see if there are issues there...

[lobosrul]

So far I've had good results from the Thecus N5200 (raid-5 mode). I've tested it with 3 concurrent ~4mbs x264 streams at once and they all played smoothly. I did have some stuttering until I set haali splitter to give a nice big buffer. Just don't expect to be able to do that while it's writing. In other words, its a no-go for a Sage recording drive.

[mikesm]

Originally Posted by lobosrul

So far I've had good results from the Thecus N5200 (raid-5 mode). I've tested it with 3 concurrent ~4mbs x264 streams at once and they all played smoothly. I did have some stuttering until I set haali splitter to give a nice big buffer. Just don't expect to be able to do that while it's writing. In other words, its a no-go for a Sage recording drive.

That's not true at all for me. I have had 4-5 recordings going on simultaneously and a couple being played back just fine with my linux based NAS, but it is a full blown system, and not an appliance.

[lobosrul]

Originally Posted by mikesm

That's not true at all for me. I have had 4-5 recordings going on simultaneously and a couple being played back just fine with my linux based NAS, but it is a full blown system, and not an appliance.

I only meant with that specific device (N5200). Although it might actually do ok writing one or two hd streams. Thats different than dumping a file on it.

[bcjenkins]

Originally Posted by mikesm

I think you'd be better off running them all in RAID5, but even so 6 of them should be plenty fast. That should not be the source of any recording problems. Are you running Sage on the same hardware, or is it going over the network?

Did you try modifying the diskopt.sh script for your setup and running that? It helped quite a bit with filesystem performance, as the disk params are not tuned for raid by default.

Also, might you be using segate or maxtor disks? If so, you need to make sure the disks don't have the dreaded AAK firmware in them which kills raid performance, and that they are all the same make and model (at least pretty much). Heterogeneous disks in RAID is a nono, though turning off NCQ as the diskopt.sh script does help.

In any case, while those disk I/O numbers while slower than they could be, are still plenty fine for a Sage system to record and playback on. If you are running Sage over a network to this system, I would look at your network setup and see if there are issues there...

Here's the disk info:

[ 4.360022] ata3: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[ 4.370010] ata3.00: HPA detected: current 976771055, native 976773168
[ 4.370010] ata3.00: ATA-7: ST3500630AS, 3.AAK, max UDMA/133
[ 4.370010] ata3.00: 976771055 sectors, multi 0: LBA48 NCQ (depth 31/32)
[ 1.655100] ata3.00: configured for UDMA/133
[ 5.560040] ata4: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[ 5.607227] ata4.00: ATA-7: ST3500630AS, 3.AAK, max UDMA/133
[ 5.607231] ata4.00: 976773168 sectors, multi 0: LBA48 NCQ (depth 31/32)
[ 5.640025] ata4.00: configured for UDMA/133
[ 5.950036] ata5: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[ 4.650011] ata5.00: ATA-7: ST3500630AS, 3.AAK, max UDMA/133
[ 4.650011] ata5.00: 976773168 sectors, multi 0: LBA48 NCQ (depth 31/32)
[ 4.650011] ata5.00: configured for UDMA/133
[ 6.280035] ata6: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[ 4.700011] ata6.00: ATA-8: ST3500320AS, SD15, max UDMA/133
[ 4.700011] ata6.00: 976773168 sectors, multi 0: LBA48 NCQ (depth 31/32)
[ 2.110018] ata6.00: configured for UDMA/133
[ 2.250027] ata7: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[ 1.480012] ata7.00: ATA-8: ST3500320AS, SD15, max UDMA/133
[ 1.480012] ata7.00: 976773168 sectors, multi 0: LBA48 NCQ (depth 31/32)
[ 4.700011] ata7.00: configured for UDMA/133
[ 6.610035] ata8: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[ 6.611368] ata8.00: ATA-8: ST3500320AS, SD15, max UDMA/133
[ 6.611371] ata8.00: 976773168 sectors, multi 0: LBA48 NCQ (depth 31/32)
[ 1.630012] ata8.00: configured for UDMA/133

3 are older 16mb cache and 3 are newer 32mb.

I am using a modified diskopt script and the hdparm numbers were run after the script ran. It boosted the raid 0 numbers but not the raid5.

B

#!/bin/bash
#
### BEGIN INIT INFO
# Description:                  Custom disk optimization script
### END INIT INFO

. /lib/lsb/init-functions

case "$1" in
        start)


echo "Optimizing RAID Arrays..."


# This step must come first.
# See: http://www.3ware.com/KB/article.aspx?id=11050

echo "Setting max_sectors_kb to 128K for RAID5 arrays..."
for i in sda sdb sdc sdd sde sdf sdg
do
echo "Setting /dev/$i to 128K..."
echo 128 > /sys/block/"$i"/queue/max_sectors_kb
done

echo "Setting read-ahead to 64MB for /dev/md[0-1]"
blockdev --setra 65536 /dev/md0
blockdev --setra 65536 /dev/md1

echo "Setting stripe_cache_size to 16MB for /dev/md[0-1]"
echo 16384      > /sys/block/md0/md/stripe_cache_size

# if you use more than the default 64kb stripe with raid5
# this feature is broken so you need to limit it to 30MB/s
# neil has a patch, not sure when it will be merged.
echo "Setting minimum and maximum resync speed to 30MB/s..."
echo 30000 > /sys/block/md0/md/sync_speed_min
echo 30000 > /sys/block/md0/md/sync_speed_max

# Disable NCQ.
echo "Disabling NCQ..."
for i in sda sdb sdc sdd sde sdf sdg
do
echo "Disabling NCQ on $i"
echo 1 > /sys/block/"$i"/device/queue_depth
done

        ;;
        stop)
        # skip or put your own commands here
        ;;
        status)
echo "Getting max_sectors_kb from disks..."
for i in sda sdb sdc sdd sde sdf sdg
do
echo -n "$i: "
cat /sys/block/"$i"/queue/max_sectors_kb
done

echo "Getting read-ahead setting from /dev/md[0-2]"
echo -n "md0: "
blockdev --getra /dev/md0
echo -n "md1: "
blockdev --getra /dev/md1

echo "Getting stripe_cache_size from /dev/md[0-2]"
echo -n "md0: "
cat /sys/block/md0/md/stripe_cache_size

echo "Getting minimum and maximum resync speed from /dev/md[0-2]..."
echo -n "md0-min: "
cat /sys/block/md0/md/sync_speed_min
echo -n "md0-max: "
cat /sys/block/md0/md/sync_speed_max


echo "Getting NCQ settings..."
for i in sda sdb sdc sdd sde sdf sdg
do
echo -n "$i: "
cat /sys/block/"$i"/device/queue_depth
done

        ;;
esac

This is the same server as the sagetv encoders. What kind of numbers should I be seeing?

[mikesm]

Originally Posted by bcjenkins

Here's the disk info:

3 are older 16mb cache and 3 are newer 32mb.

I am using a modified diskopt script and the hdparm numbers were run after the script ran. It boosted the raid 0 numbers but not the raid5.

B

#!/bin/bash
#
### BEGIN INIT INFO
# Description:                  Custom disk optimization script
### END INIT INFO

. /lib/lsb/init-functions

case "$1" in
        start)


echo "Optimizing RAID Arrays..."


# This step must come first.
# See: http://www.3ware.com/KB/article.aspx?id=11050

echo "Setting max_sectors_kb to 128K for RAID5 arrays..."
for i in sda sdb sdc sdd sde sdf sdg
do
echo "Setting /dev/$i to 128K..."
echo 128 > /sys/block/"$i"/queue/max_sectors_kb
done

echo "Setting read-ahead to 64MB for /dev/md[0-1]"
blockdev --setra 65536 /dev/md0
blockdev --setra 65536 /dev/md1

echo "Setting stripe_cache_size to 16MB for /dev/md[0-1]"
echo 16384      > /sys/block/md0/md/stripe_cache_size

# if you use more than the default 64kb stripe with raid5
# this feature is broken so you need to limit it to 30MB/s
# neil has a patch, not sure when it will be merged.
echo "Setting minimum and maximum resync speed to 30MB/s..."
echo 30000 > /sys/block/md0/md/sync_speed_min
echo 30000 > /sys/block/md0/md/sync_speed_max

# Disable NCQ.
echo "Disabling NCQ..."
for i in sda sdb sdc sdd sde sdf sdg
do
echo "Disabling NCQ on $i"
echo 1 > /sys/block/"$i"/device/queue_depth
done

        ;;
        stop)
        # skip or put your own commands here
        ;;
        status)
echo "Getting max_sectors_kb from disks..."
for i in sda sdb sdc sdd sde sdf sdg
do
echo -n "$i: "
cat /sys/block/"$i"/queue/max_sectors_kb
done

echo "Getting read-ahead setting from /dev/md[0-2]"
echo -n "md0: "
blockdev --getra /dev/md0
echo -n "md1: "
blockdev --getra /dev/md1

echo "Getting stripe_cache_size from /dev/md[0-2]"
echo -n "md0: "
cat /sys/block/md0/md/stripe_cache_size

echo "Getting minimum and maximum resync speed from /dev/md[0-2]..."
echo -n "md0-min: "
cat /sys/block/md0/md/sync_speed_min
echo -n "md0-max: "
cat /sys/block/md0/md/sync_speed_max


echo "Getting NCQ settings..."
for i in sda sdb sdc sdd sde sdf sdg
do
echo -n "$i: "
cat /sys/block/"$i"/device/queue_depth
done

        ;;
esac

This is the same server as the sagetv encoders. What kind of numbers should I be seeing?

EEK! Your Seagate drives have the dreaded AAK firmware! This was discussed earlier in this thread. There is a big performance problem with the AAK firmware, and it seems to especially hurt in RAID. See this link: http://www.fluffles.net/articles/seagate-AAK-firmware

You should get these RMA'd for a version with different firmware in them. I think Kjake had the same issue and also suffered in RAID performance.

Maybe you can get them all upgraded to the same model drives. RAID likes everything being the same for optimum performance, but in any case, you need to get those AAK's out of there if you are going to see decent numbers. Seagate seems willing to swap them out, but you will need to make sure you get a part number in return that is different, maybe the part number of the other 500GB drives...

[bcjenkins]

From: http://www.fluffles.net/forum/storage/34.3
Firmware update for the 500gb aaks.. http://www.mediafire.com/?30ddatf19xs

Now hdparm reports:

hdparm -tT /dev/mapper/video_vg-video_lv

/dev/mapper/video_vg-video_lv:
Timing cached reads: 8842 MB in 2.00 seconds = 4425.90 MB/sec
Timing buffered disk reads: 452 MB in 3.01 seconds = 150.13 MB/sec

What speeds should I see?

B

PS - The update seems to have made an improvement.

[mikesm]

Originally Posted by bcjenkins

From: http://www.fluffles.net/forum/storage/34.3
Firmware update for the 500gb aaks.. http://www.mediafire.com/?30ddatf19xs

Now hdparm reports:

hdparm -tT /dev/mapper/video_vg-video_lv

/dev/mapper/video_vg-video_lv:
Timing cached reads: 8842 MB in 2.00 seconds = 4425.90 MB/sec
Timing buffered disk reads: 452 MB in 3.01 seconds = 150.13 MB/sec

What speeds should I see?

B

PS - The update seems to have made an improvement.

For 6 disks, you should be seeing double that. You are running XFS, and the stripe size on the disks is 128KB or 256KB? Is the xfs volume very fragmented?

hdparam isn't the best tool by the way. Try running bonnie++ and see what it says...

[mikesm]

BTW, Fry's is now running a special - a retail maxtor 1TB SATA II disk for $199, no rebate. This is a decent disk for raid5, though I still prefer my hitachi's...

[toricred]

Originally Posted by mikesm

BTW, Fry's is now running a special - a retail maxtor 1TB SATA II disk for $199, no rebate. This is a decent disk for raid5, though I still prefer my hitachi's...

Do you have a link for that?

[mikesm]

Originally Posted by toricred

Do you have a link for that?

Sure: http://latimes.p2ionline.com/shoppin...21_5362617.jpg

BTW, they also have the Q6600 for $189 as well... :-)

[toricred]

I was afraid of that. It's in store only. The nearest Fry's is about 1000 miles away. I really hate living in a state that is so backwards.

[bcjenkins]

Originally Posted by toricred

I was afraid of that. It's in store only. The nearest Fry's is about 1000 miles away. I really hate living in a state that is so backwards.

But you have green chile.