ASUS P7P55D-E ..... Testing level 1 Passed

[erikatcuse]

4/16/10 Update: Disk Rebuild and a subsequent parity check both passed.  Along with a 2TB drive preclear.  The issue seemed to be memory seated improperly or disabling disk spindown and spinup groups.  I'll re-enable spindown and start another parity check sometime this weekend to see if it locksup but right now this board passes.

 

4/14/10 Update:  Preclear has almost completed post read and no problems as of yet.  It looks to have been a memory problem.  Either seated incorrectly or Bios had the timings set to low.  I'll report back after I add the new drive to the array.

 

 

 

I bought the ASUS P7P55D-E 9 SATA ports.  I've filled 8 of them with a combo of 1 & 2 TB drives.  It has an i3 processor and 4gb of memory in it.

 

I ran into a bit of a problem at first with the flash drive.  The serial number wasn't being reported correctly.  I changed the USB setting in the bios to Force FDD and all was good.

 

I'm running a parity check right now and its reporting 70MB/S.  So far it looks like a good board.  I'll do some transfer tests later tonight and check the HD speeds as well.  I'm also going to be upgrading one of the 1TB drives to a 2TB drive as well since my 2 1430sa controllers are back ordered.

[Msan]

70KB/s ? I hope that's a typo Looks better now

[bman]

Curious to learn more about your experiences with this new board of yours as you learn yourself.  I am on a first-gen P4 platform with only 4 ports onboard, and an overly-saturated PCI bus to handle the other 6 drives.

 

Parity checks take me many days and the whole thing is only 5TB in size... so you could say I can somewhat easily be persuaded to upgrade if I knew what worked really well.

 

[erikatcuse]

Parity Check finished in 8h 20m that's with 6 drives + parity.  2 2TB and 4 1TB drives.  I'm also preclearing a new 2tb drive as well.

 

I did flash the bios and had to change the power settings.  The array kept dropping off line.  After the flash and changing the S3+S1 to only S3 No problems.

 

I'm going to run another parity check to see if it picks up any errors.

 

Overall a very nice board

[prostuff1]

When we get a syslog attached and I can look through it I will update/add this board to the Hardware Compatibility Page.

[erikatcuse]

When I got up this morning I checked on the second parity check and the server was unresponsive.  Nothing on the monitor, I couldn't ping it or access the web interface.  When I checked my router software and looked at the list of connected devices the unraid server wasn't listed. 

 

 

I ran memtest this morning and it passed.  I have the syslog from yesterday and today and can't find any glaring errors.  I went back into the bios and disabled acpi but left s3 on for now.

syslog-2010-04-09.txt

syslog-2010-04-10.txt

[erikatcuse]

Unraid locked up again  .  I'm not sure what's going on but it seems random.  This time a parity check was running and I was attempting to clear a 2tb drive.  I've decided to just let the parity check that started when I rebooted complete before I do anything else.  I also went into BIOS and found everything possible to disable and did.  Latest log file is attached.  Any suggestions would be appreciated.

 

Thanks

 

Erik

syslog-2010-04-102.txt

[erikatcuse]

Parity check completed with 34 errors...I'm going to run it again.  Here is a continuation of the syslog from last night to this morning.

syslog-2010-04-11.txt

[erikatcuse]

Parity check 3 done with no errors and is running fine.  I'm going to try and preclear a new disk I have and see if it locks up again.

 

 

[erikatcuse]

Tried another preclear and it froze up.  So I decided to add the drive to the array to see if it would clear that way.  No such luck it froze almost instantly.  The good news is I was able to capture the last info on the screen with a camera.

 

I've run a SMART report on the drive in question and its passed.

 

I've also changed the SATA cable even though it's brand new to see if that could be the problem.

 

If anyone has any other suggestions I'm open.

 

 

[purko]

If anyone has any other suggestions I'm open.

You said you flashed your BIOS.  After you did that, did you verify that the voltages and the timings for your memory are set correctly?

 

[erikatcuse]

I currently have it all set to Auto but I've look at the bios and they look to be correct.

[SSD]

Check the power connection to the drive you were having problems preclearing. If a drive loses power it can cause machine crashes. I believe Joe L had a similar problem that he isolated to a faulty splitter that caused random crashes. Try a spare power connector from the PSU bypassing any backplane / drive cages.

[erikatcuse]

I changes the cables around but I don't have a backplane or any splitters.  I also pulled and reseated the memory.  I did change the memory settings to be more conservative then the auto settings.  I'm now in the process of trying the pre clear script again.  I had this morning but the computer rebooted at about the 3 hour mark.  If it locks up again I'm going to put a 1 TB drive that I have and try to preclear it.  I have no indications the drive is bad but you never know. 

[purko]

I also pulled and reseated the memory.   I did change the memory settings to be more conservative then the auto settings.

Any time you reseat/replace memory or change memory settings, first thing you do is run memtest for a few hours.

 

[erikatcuse]

Looks like reseating the memory and/or changing the auto timings in the bios did it.  The bios was a bit aggressive with the timings so I changed them to manufacturer advertised specs.

 

The good news is the preclear is 37% through the post read on my 2 TB drive and all is well.  Once the preclear is done we'll see how adding it to the array goes.  After that I'll rerun the parity check and report back.  Thanks everyone for the help.

[MikeL]

 

 

  I was going to suggest checking the mem setting in BIOS. When set to auto, the BIOS gets it wrong a lot of the time. Good to see you may have it worked out, as I will be keeping an eye on you reporting back. I am looking for a new board, and this one caught my eye.

Thanks for the updates.

 

Looks like reseating the memory and/or changing the auto timings in the bios did it.  The bios was a bit aggressive with the timings so I changed them to manufacturer advertised specs.

 

The good news is the preclear is 37% through the post read on my 2 TB drive and all is well.  Once the preclear is done we'll see how adding it to the array goes.  After that I'll rerun the parity check and report back.  Thanks everyone for the help.

[erikatcuse]

Preclear passed...I decided to simulate a disk fail and took a drive with about 50gb on it off line.  Restarted the array and it worked as advertised.  I was able to access the files and use them.

 

I then assigned my new drive to the simulated failed slot and am rebuilding parity right now.  The build is currently running about 70,000KB/sec.  I've marked the start time and will get an avg after it completes.

 

 

 

 

[Rajahal]

Preclear passed...I decided to simulate a disk fail and took a drive with about 50gb on it off line.  Restarted the array and it worked as advertised.  I was able to access the files and use them.

 

I then assigned my new drive to the simulated failed slot and am rebuilding parity right now.  The build is currently running about 70,000KB/sec.  I've marked the start time and will get an avg after it completes.

 

 

 

 

 

Hopefully you just misspoke, but if you are rebuilding parity (i.e. you pressed the 'restore' button), then you just lost that 50 GB of data.  If you are rebuilding the failed drive (i.e. you pressed the 'start' button) then all is well.

[purko]

I then assigned my new drive to the simulated failed slot and am rebuilding parity right now.

Hopefully you just misspoke, but if you are rebuilding parity (i.e. you pressed the 'restore' button), then you just lost that 50 GB of data.

I am sure that he misspoke, or he wouldn't be rebuilding anything!   

BTW, when we think about it, rebuilding a failed disk and building parity are basically the same thing under the hood:

You add up the same bit from each of the other disks, and then you write to the disk in question either 0 or 1 to make the total summ be even.

 

[erikatcuse]

I'm on my second attempt.  The server was locked up this morning again no screen to see and no syslog so I rebooted.  I decided to do a mem test on 1 stick of memory again and it passed several times.  Tonight I started another data-rebuild on the new 2tb disk.  I checked it at about 17% and it was still up.  I was looking reading another thread in the 4.4 section and it was mentioned that some people were having issues with spindown delay and spinup groups.  So just in case I've disabled them.  

 

It hit 50% and the 5 1TB drives are finished.  So far 4h 33m.  The parity speed also jumped from 40MB/s to over 100MB/s but that's to be expected.  The 2TB data drive only has about 40gb of data on it.

[erikatcuse]

Completed disk build in 7h 50m.  I'm doing a parity check now as well as preclearing the replaced drive.

[erikatcuse]

From the syslog the rebuild took time=28246sec rate=69160K/sec

 

This was replacing a 2TB drive (Had the least amount of data on it) with a 2TB (precleared) drive.  I don't thing it matters since the whole disk was written to but there was 67gb of data on the drive.

 

I then ran a parity check it completed with no errors and from the syslog too time=27841sec rate=70166K/sec to complete

 

At the same time as the parity check I started a preclear of the old 2TB drive.  It's 9 hours into it and no problems so far.

 

So my problems are resolved and it looks like it the memory could have been seated poorly, or disabling autospin and spinup groups.

 

Despite the problems I like the board and fact that I have 8 ports on-board.  I have 2 Adaptec 1430SA coming next week from Walmart.  Ill add them and change the drive configuration around to take advantage of all the PCIe lanes.  

 

Unless I have any other problems this board finally passes.

[erikatcuse]

Added 2 Adaptec 1430sa to the server yesterday and reconfigured the hard drives.  In total I have 8 drives, 5 1tb (mixture of WD, Hitachi, Samsung) and 3 2tb (Hitachi, Seagate) 2 on each controller.  Parity finish fine and I've re-enabled diskspindown.  From the sys log sync done. time=28608sec rate=68285K/sec.

 

Hard drive speed test from unmenu

 

/dev/sda:  (WD green 1tb)

Timing cached reads:  10294 MB in  2.00 seconds = 5153.88 MB/sec

Timing buffered disk reads:  236 MB in  3.01 seconds =  78.53 MB/sec

 

/dev/sdb: (Hitachi 2tb)

Timing cached reads:  10168 MB in  2.00 seconds = 5090.75 MB/sec

Timing buffered disk reads:  362 MB in  3.01 seconds = 120.09 MB/sec

 

/dev/sdc: (WD green 1tb)

Timing cached reads:  10168 MB in  2.00 seconds = 5090.72 MB/sec

Timing buffered disk reads:  250 MB in  3.02 seconds =  82.87 MB/sec

 

/dev/sdd: (Samsung 1tb)

Timing cached reads:  10756 MB in  2.00 seconds = 5385.45 MB/sec

Timing buffered disk reads:  284 MB in  3.00 seconds =  94.62 MB/sec

 

/dev/sde: (Hitachi 2tb)

Timing cached reads:  10100 MB in  2.00 seconds = 5056.42 MB/sec

Timing buffered disk reads:  356 MB in  3.01 seconds = 118.19 MB/sec

 

/dev/sdf: (Hitachi 1tb)

Timing cached reads:  9946 MB in  2.00 seconds = 4978.84 MB/sec

Timing buffered disk reads:  248 MB in  3.00 seconds =  82.61 MB/sec

 

/dev/sdh: (Seagate 2tb)

Timing cached reads:  9894 MB in  2.00 seconds = 4953.51 MB/sec

Timing buffered disk reads:  346 MB in  3.01 seconds = 115.10 MB/sec

 

/dev/sdi: (Hitachi 1tb)

Timing cached reads:  10114 MB in  2.00 seconds = 5063.46 MB/sec

Timing buffered disk reads:  250 MB in  3.01 seconds =  82.94 MB/sec

 

[DaveK]

Hi All, was there any more progress on using the ASUS P7P55D-E motherboard for UNRAID? A friend and I want to use this motherboard and we were wondering on the status of level 2.