Hi guys,

 

I have been running 4.5.6 for a few months now and tonight I started using an unused disk in my system (disk2 of 14 disks). I started to write 3 movie.MKV files to the disk and part way through the second file, the disk suddenly became "read only" (I can not write to the disk even using Midnight Commander). I rebooted the server, but disk2 was still "read only", so I used Putty to Telnet in, unmount disk2 and then run reiserfsck to check on the condition of the disk. Reiser has now been running (I think) for about 4 hours and is not reporting anything back to me. The last message I got was "ReiserFS journal '/dev/md2' in blocks [18..8211]: 2 transactions replayed" and then I have a non-blinking cursor on the next line. I think the system is locked up, but I don't know for sure.

 

I tried opening another Putty window and all I get is a cursor with no normal prompt.

 

My question - How do I shut down safely? I don't want to just pull out the power cord. Is there some way to regain control in a situation like this?

 

 

The web management page perhaps?

Hi Joe!

 

I was following the procedure here:

 

http://lime-technology.com/wiki/index.php?title=Check_Disk_Filesystems

 

so one of the first things I did was to stop Samba, though I don't know if Samba has anything to do with the web interface. At any rate I can no longer access the server from the web.

Hi Joe!

 

I was following the procedure here:

 

http://lime-technology.com/wiki/index.php?title=Check_Disk_Filesystems

 

so one of the first things I did was to stop Samba, though I don't know if Samba has anything to do with the web interface. At any rate I can no longer access the server from the web.

No, it has nothing to do with the web-interface.  It will prevent you from seeing the shared drives on the LAN, but that is an entirely different service than the web-management page.

Ok, I just got back and there were no changes since last night, so I assume that reiserfsck locked up the entire system as soon as it started checking disk2. Since there seemed to be no better way to shut down the system, I shut it down by using the switch on the power supply. While it was shut down I reattached the SATA cable going to disk2 just in case it is/was the cause of the original problem. I started the system back up and everything seems to be running fine...I can now delete and write files to disk2 again, though I am not going to do much right now, as the system has initiated a parity-check on its own - I assume it was because of the abnormal shut down.

 

This brings up a question....what happens in the event of a power failure? The system will obviously not be shut down correctly and safely. Maybe some sort of information as to how the system will react and what steps need to be taken after an abnormal shutdown should be included in the unRAID FAQ...unless it is already there and I missed it. I didn't find any info on what to do on a total lockup and/or abnormal shutdown and I think it will happen to just about everyone sooner or later.

 

So my plan right now is to let the system do a complete parity check, as it was just about due for its monthly check anyway.

 

When I first built this sytem, I had 2 WD20EARS drives being used as disk1 and disk2 and I had a lot of problems reading and writing to disk2. My assumption at the time (which I now think is incorrect) was that the problem was with the "EARS" drives, so I replaced them with Samsung F4's, drives which I like a lot better. Since that time I have filled up disk1 (no problems reading or writing) and when I had this problem last night it was the first time had I tried using disk2 since I changed the brand of hard drive. Since I now have had a problem which is very similar, if not identical, to my problem with the "EARS" drive in the same slot, my new conclusion is that I have some other hardware problem in my system that is wreaking havoc with disk2 regardless of what drive is attached. Does this seem reasonable? The controller is an 8 port Supermicro unit using 2 SAS-SATA breakout cables, so I can not change either the cable or controller easily. I can forfeit my cache drive to use its cable (separate SATA) and controller (from motherboard)....Is this a good way to troubleshoot the problem?

 

This brings up a question....what happens in the event of a power failure? The system will obviously not be shut down correctly and safely. Maybe some sort of information as to how the system will react and what steps need to be taken after an abnormal shutdown should be included in the unRAID FAQ...unless it is already there and I missed it. I didn't find any info on what to do on a total lockup and/or abnormal shutdown and I think it will happen to just about everyone sooner or later.

 

After a power failure, unRAID will perform an automatic parity check on restart.  If this identifies a problem, then normal recovery process should be invoked.  This is why many unRAID users have a smart UPS powering their server.  With an average of one power outage a day, I'd be foolish not to use a UPS.

 

So my plan right now is to let the system do a complete parity check, as it was just about due for its monthly check anyway.

 

Unless you are absolutely sure that there was no write activity on your array at the time of the failure, you would be well advised to always allow the parity check to complete.

 

When I first built this sytem, I had 2 WD20EARS drives being used as disk1 and disk2 and I had a lot of problems reading and writing to disk2. My assumption at the time (which I now think is incorrect) was that the problem was with the "EARS" drives, so I replaced them with Samsung F4's, drives which I like a lot better.

 

Many of us are using EARS drives without problem ... indeed, some prefer the EARS over other manufacturers' offerings.  Unless the firmware bugs have been fixed recently, you should be very cautious about using the F4 drives!

 

Since that time I have filled up disk1 (no problems reading or writing) and when I had this problem last night it was the first time had I tried using disk2 since I changed the brand of hard drive. Since I now have had a problem which is very similar, if not identical, to my problem with the "EARS" drive in the same slot, my new conclusion is that I have some other hardware problem in my system that is wreaking havoc with disk2 regardless of what drive is attached. Does this seem reasonable?

 

That seems entirely logical.

 

 

The controller is an 8 port Supermicro unit using 2 SAS-SATA breakout cables, so I can not change either the cable or controller easily. I can forfeit my cache drive to use its cable (separate SATA) and controller (from motherboard)....Is this a good way to troubleshoot the problem?

 

I'm not familiar with the Supermicro controller but is it not possible to use an alternate port of the eight available?

 

Doesn't your mobo have more than one SATA port?  Can you not use another mobo port (buy another SATA cable, if you have to).

Thanks for the reply, PeterB!

Unless the firmware bugs have been fixed recently, you should be very cautious about using the F4 drives!

Hmmm...what firmware bugs? I have 9 of the Samsung F4 2TB drives in my array. They run incredibly cool and have not given me even the slightest problem. What issues are people having with these drives?

I'm not familiar with the Supermicro controller but is it not possible to use an alternate port of the eight available?

 

Doesn't your mobo have more than one SATA port?  Can you not use another mobo port (buy another SATA cable, if you have to).

I have 8 SATA ports on my mobo and 8 on the Supermicro controller. My array consists of 14 2TB data drives, one 2TB parity drive, and one 1TB cache drive, so I have no controllers to spare. I will be able to play with this more in a few weeks, as I am upgrading to a 24 disk system (22 data drives, 1 parity, and 1 cache) and I will have 30 SATA controllers total available (6 for troubleshooting).

See this link on F4 issues and firmware update:

 

http://lime-technology.com/forum/index.php?topic=9339.0

Hmmm...what firmware bugs? I have 9 of the Samsung F4 2TB drives in my array. They run incredibly cool and have not given me even the slightest problem. What issues are people having with these drives?

 

Non-reported and silent data corruption issues under certain conditions. I highly suggest you apply the firmware patches immediately to all your Samsung F4 drives if you value your data. Once again, the thread with all the details is here: http://lime-technology.com/forum/index.php?topic=9339.0

Thanks, guys! Wow, I was totally unaware of the problem. I stopped my parity check (it had reported 63 errors, all of which seem to be on disk14, one of the Samsung disks) and restarted a read only parity check using:

 

/root/mdcmd/ check NOCORRECT

 

from a Putty window.

 

I plan on updating the firmware by creating a DOS boot flash drive, but the link in the thread is broken. I am going to do a Google search for the "how to" on setting up a DOS boot flash drive that supports SATA, but if someone already knows of a tutorial, please let me know here.

 

Many thanks again!