Supermicro AOC-SASLP-MV8 and disk drop outs

S

shaun.arundell

n00b
Joined
Mar 11, 2010
Messages
3
I have recently put together a new windows 7 file server based on a gigabyte GA-H55N-USB3 and the Supermicro AOC-SASLP-MV8. I have 7 2T F4s and 1 WD .5TB disk attached. The WD is via a MB SATA port and the 7 2T F4's off the AOC-SASLP-MV8 via 2 SAN- 4 SATA breakout cables.

The AOC-SASLP-MV8 is running the latest firmware

I am getting random disk drop outs on the AOC-SASLP-MV8 usually when I am putting a high write load onto one of the disks. Thought its hard to pick a pattern, because it only happen randomly.

The disk that drops out disappears from windows explorer and device manager, but all other disks continue to function. A reboot will bring the disk back.

Here a 2 examples of when it has happened

1. If I copy from an external share onto the a disk (about 50MB/Sec) it will usually will usually cause a drop out.

2. Sometime when calculating and writing raid parity data with flexraid, the parity disk being written will drop out.

But I have not been able to get a disk to drop out when

a. copy from the MB sata port to a disk on the AOC-SASLP-MV8 (120M/sec)
b. copy from one disk on the AOC-SASLP-MV8 to another on the AOC-SASLP-MV8 (120MB/sec)

Its very strange because I have another file server running WHS with a mix of 2TB 3.5 WD greens and .75TB WD scopio blues and the same model AOC-SASLP-MV8 and have had no drops outs.

I thought if might be heat, as I have seen hints of this behavior in various posts on the net. So I moved it all to another case, with lots of fans - no difference - but its hard to tell what temp the AOC-SASLP-MV8 is running at.

Any ideas ?
Do I give up on the AOC-SASLP-MV8 and get another type of 8 port SATA board ?
Another case ?
Put a fan directly blowing onto the marvel chip on the board ?.
Could it be the F4s ?

I am stumped on how to track down the problem

thanks in advance
Shaun
 
I am also having this problem. I updated to the latest firmware, and it still happens. I haven't done as much testing as you to say it does/doesn't happen with different port combinations, all I can confirm is that a drive will drop out under a high write load. It drops out of the data pool, and th event viewer will show an error that says \Device\Scsi\mv64xx1 did not respond. In MY case, a reboot isn't sufficient. I have to actually power down the server to reset it (if I don't power down, I can see in the SASLP bios boot that there's still a problem with that drive). It's not always the same drive, but it's always a drive on the SASLP. I also have a couple of F4s on this card, but I've also seen it happen on an F3. I can't say for sure if I've seen it move to a non-Samsung.
 
I am having the exact same problem. I am suspecting either the card or the inablity to work with the ears drives. I just went through the process of changing the default timer (see this article http://www.mediasmartserver.net/201...-format-hard-drives-with-windows-home-server/)

thinking this may be the cause but it is still ocurring. Drives connected to this card are timing out and dropping (go missing) from the WHS drive pool). This is becoming epidemic and am struggling to troubleshoot. THere is no newer microcode to try. So there are a couple of steps i am going to take. I have opened an incident with WDs (no response yet) and next is with Supermicro. THe only other thing i can think of is to take the ears drives off of the sata card and put them directly on the motherboard and move the eads drives to the card.

Have you fixed your issue. This is very frustrating.!!!
 
I've got sick of trying to solve this problem. So I bought a new MB, a gigabtype microatx H55M USB3. 7 useable SATA ports on board. Anded add a 2 port SATA2 II PCIE card for 9 ports total. It took me < 1hour to install the new motherboard, reinstall windows 7 and reconnect all my disk (8*2T + 1 .5TB from windows).

Its the best thing I ever did. It works fast, perfect, no drop out, not issues with slow boots, with delays for disk spin ups. I'll leave the SAS cards for the guys doing 10+ disks from now on.

also

The new MB and PCIE card cost a lot less then the super micro card and the 2 SAN-sata cables.

So its cheaper
 
thanks for the update. I dont want to have to the same thing. This is a WHS system with about 8 TB all on a mixture of WD eads and ears drives (ears being the advanced format). Would rather not have to re-install WHS unless the last resort.

I'm sure its related to these new advanced format hard drives re: 4096 vs 512 sector size. I've just emailed supermicro (lot of good it will do i'm sure). but maybe they are aware of some issue. I've run the card for over a year now with no issues. Since i started adding ears to this card its been nothing but trouble. As mentioned next step is to remove ears from this card and see what happens.
 
I've had similiar issues for couple of times (in two years) but nothing major and simple cold restart fixed it. It could be card is overheating (I installed tiny 40mm fan on it's heatsink). Could be cables. Maybe power flactuates.
 
You guys are probably using a badly aligned partition so you get very poor performance on EARS drives. But i wonder, have you checked the SMART data? Disks dropping from conventional RAIDs sounds more like uBER to me. You can check the raw data value of:

Reallocated Sector Count = fixed bad sectors; could have been a problem in the past but no longer cause problems
Current Pending Sector = immediate danger, active bad sector, beware!
UDMA CRC Error Count = cable problem

Overheating wouldn't be my guess. And if you have too few power you would see this with spinup and the system would shut down due to overcurrent protection. Once the system is running the power draw is much lower than the spinup current required.
 
98_1LE said:
Is it possible you are running out of +12v power to the drives?
Click to expand...

This happened to me several years ago. Drives randomly dropping, reboot wouldn't help, etc. It turned out to be a failing Seasonic power supply. I replaced it and the problems vanished.

Though in this case with a number of different people reporting the problem, it seems less likely.
 
sub.mesa said:
You guys are probably using a badly aligned partition so you get very poor performance on EARS drives. But i wonder, have you checked the SMART data? Disks dropping from conventional RAIDs sounds more like uBER to me. You can check the raw data value of:....
Click to expand...

I'm not even running mine in RAID, and I only have a couple advanced format drives. I've seen many other reports of this, so it appears to be some glitch with the SASLP.

I had contaced Supermicro, and they sent me the latest firmware update (which is actually from some time last year), and I THOUGHT that it fixed the issue. But then it happened to me two times in a row last night as I was trying to rip a movie. It happened, rebooted, deleted the foler, tried to rip again and it happened again. I am on WHS, and I use the Disk Management add-in. I recently started disabling the monitoring feature where it's constantly updating the drive temp and bandwidth. But I had turned that back on yesterday. After the two failed attemps, I turned that back off, and the movie ripped fine. It may be coincidence, but perhaps it could be some incompatibilitie with the SASLP handling high bandwidth data throuput and those SMART calls simultaneously?

FWIW, I know the F4 firmware update was due to corruption during certain SMART checks, but I've applied that update, so this shouldn't be related. And as I said, I've also seen it happen to F3 drives on the SASLP.
 
Can you run HDTune and select your disk on that controller? Do you see the SMART data in Health tab?

I'm guessing that if you don't, this controller is working on a SCSI layer instead. You may also check that with AS SSD, which lists the driver in use. I'm guessing that's no ATA or AHCI driver, but a Marvell specific driver that also adds RAID functions.
 
rmd3003 said:
I've had similiar issues for couple of times (in two years) but nothing major and simple cold restart fixed it. It could be card is overheating (I installed tiny 40mm fan on it's heatsink). Could be cables. Maybe power flactuates.
Click to expand...

This probably hits the nail in the head. I used to have similar issues with my card until i rigged a fan to blow air directly at the heatsink. Solved my problems. If nothing else, give it a try.
 
i'm having trouble believing that its heat related (in my case at least). This has worked fine for over a year and wasnt having any issues until adding EARS drives to this controller. I'm also on WHS and am running smart add-in and it reports no issues with the drives. Its also not an alignment issue as they are jumpered (across pin 7/8) and are in the data pool only, not used as the system drive. There's a number of us reporting issues with this card and (I think) advanced format drives.
 
sub.mesa said:
Can you run HDTune and select your disk on that controller? Do you see the SMART data in Health tab?
Click to expand...

I can select the SASLP connected drives in the HDTune drop-down, but HDTune doesn't show anything in the health box for ANY of my drives, even those connected to MB ports. But I do get smart data from other tools, for both my MB connected drives and the SASLP connected drives (but NOT on the ones connected to the SiI3124 controller w/PMs).
 
dscline said:
it happened to me two times in a row last night as I was trying to rip a movie. It happened, rebooted, deleted the foler, tried to rip again and it happened again. I am on WHS, and I use the Disk Management add-in. I recently started disabling the monitoring feature where it's constantly updating the drive temp and bandwidth. But I had turned that back on yesterday. After the two failed attemps, I turned that back off, and the movie ripped fine. It may be coincidence, but perhaps it could be some incompatibilitie with the SASLP handling high bandwidth data throuput and those SMART calls simultaneously?
Click to expand...

FWIW, this seems to be reproduceable for me. I disabled the monitoring utility that monitors HD temps and bandwidth, and I've ripped several movies over the past few days. Last night, I turned monitoring back on, ripped a movie, and a drive dropped out. I've not yet had a drive drop out while I've had that function disabled.

So, does anyone else who has problems with drives dropping out also use some utility that monitors temps or other SMART functions of the HD? If so, can you disable it, and confirm whether or not that is a factor in your system?
 
i am also on whs and have the data management plug-in. I have moved 2 of 3 drives off of the saslp-mv8 and back to the motherboard. I have left one on the m8v (wd15eads) and it has dropped off twice in the last 24 hrs. All i get is timeout. I dont know what has changed (other than the newer hard drives). I can try your suggestion and see if it changes anything. The reason i left the eads drive on the mv8 as i was suspecting it was ears related. Now i'm not sure what the heck is going on.
 
mayhem99 said:
I can try your suggestion and see if it changes anything. The reason i left the eads drive on the mv8 as i was suspecting it was ears related. Now i'm not sure what the heck is going on.
Click to expand...

There are plenty of reports of it happening with all kinds of different drives, so I don't think it's drive related. But I'll look forward to seeing if disabling real time monitoring in Drive Management improves things for you. I don't know why that would make a difference, but it definitely seems to in my case.
 
this is talking about a number of different issues ,not just drives dropping. It seems to be all over the place. They did provide a link for firmware update, but now i'm a little confused (seems to be faily often lately). I will check when i'm home but on boot up is says firmware is 0.3.15N I think when i check in device manager (WHS) the driver is 0.3.1.22 - so not sure if he's talking driver or firmware or maybe its supposed to match? Also i mentioned earlier in this forum that i have sent an email to SM but yet no repsonse. Been a week already. Has anyelse contacted them? thx...
 
Yes, I contacted them. They told me to update to the latest firmware (which they labeled 3.1.0.21, but the actual file was 6480.bin), but like others have reported, the issue continued after the firmware update. I let them know I was still having the issue, and this was their response:

There are many possible issue may cause this. Hard drive firmware issue, error detection and correction time out issue or driver issue. It&#8217;s very difficult to tell which one is the major issue.
Click to expand...

That's it. :rolleyes:
 
at least they responded to you. I see no option now (I've spent more time fixing ...rebooting ...missing disks than using my WHS), but to pull the card and move drives back to native sata on the MB (GA785GT-UD3H). Only issue is 6 sata ports and 7 current drives. Guess just upgrade them to greater sizes. So frustrating after running fine for over a year. I still figure its something to do with the newer eads / ears drives.
 
Has anyone been able to confirm or refute my theory that SMART calls may be involved? I've had real time performance monitoring turned off on my machine for the past week, and have probably 15-20 movie rips in that time, with not a single drop. I've had drops very frquently at the times when that was enabled. This may just be a coincidence in my setup, but if we could find a common factor in the cases where it happens, we'd have something to provide Supermicro with to track down and fix the issue.
 
You must log in or register to reply here.
Back
Top