Here's a weird one.  I've been ripping DVDs all day to the unRAID server.  I'm not sure how it started, the server seemed to drop offline.  I couldn't connect anymore so I tried to manually shut it down by pressing the power button and nothing happened, so I'm thinking it crashed.  So I shut it down by pressing and holding the power button.

  I started it up and it was fine, and I COULD connect via HTTP initially.  The Parity drive had started a check and got through about 1%.  Then when I tried to refresh the web interface, I couldn't get through.  Everything else works.  I can access the disks and shares from the iMac and Windows machine, I can get into the server via telnet.  But when I try to log on via a browser from both the Mac & Windows machine, it hangs, waiting for a response and getting none.  I've tried it with Safari, Firefox and IE.  I'm including the syslog and a few commands I noticed Joe L. suggested in a similar situation.

 

 

Any ideas?

eth_tests_7:10:10.txt

syslog.txt

The syslog shows that 4 of your disks are having difficulty.

These may be on the Jmicron ports. Are they set for IDE emulation or AHCI?

 

When they reset it will cause pauses.

As for the emhttp going away, that might be an OOM condition when you move allot of data to the machine, but I do not see that in this syslog.

 

I do see resets on 4 of the hard drives right away.

OK, I think I understand what you're saying.  The only way I know how to check if they are in IDE mode is to restart the server and go into the BIOS, is that right?

 

As far as them having problems, I did notice that when I restarted the server the first time it took quite some time for those four disks to mount, but they eventually did.  I think it was because I shut it down manually when it wouldn't shut down normally.  I assumed those four disks were just trying to get back to a point where they could mount, I've seen that happen with my other machines if they were shut down improperly.  There definitely was a lot of disk activity while they were trying to mount.

 

A couple of questions, without the HTTP interface, how can I see if the parity check is finished and if there were any errors and assuming that's alright, what would be my next move?  I'm thinking of restarting the server again, but I'm not sure.  THANKS!

A couple of questions, without the HTTP interface, how can I see if the parity check is finished and if there were any errors and assuming that's alright, what would be my next move?  I'm thinking of restarting the server again, but I'm not sure.  THANKS!

 

login as root with telnet.

 

Type the following

/root/mdcmd status | grep '^md'

 

On my idle system the output is

mdVersion=0.95.4

mdState=STARTED

mdNumProtected=15

mdNumDisabled=0

mdDisabledDisk=0

mdNumInvalid=0

mdInvalidDisk=0

mdNumMissing=0

mdMissingDisk=0

mdNumNew=0

mdResync=0

Hmmm, this is what I get...

 

 

Linux 2.6.32.9-unRAID.

root@unRAID:~# /root/mdcmd status | grep '^md'

Binary file (standard input) matches

root@unRAID:~#

 

Use this:

/root/mdcmd status | strings | grep '^md'

 

On some machines, there is a null character returned confusing "grep." The "strings" command cleans it up to readable strings only.

Something goofy is going on.  It dropped offline again.  I still couldn't get to it by http and now I couldn't get to it thru smb.  But I still could get to it with telnet.  So I tried to restart from there using "shutdown -r now" and it seemed to start to shutdown but froze.  I then I forced a shutdown by holding the power button.  

 

I just restarted it and everything seems to be back and it's running a parity check, looks like it wants about 10 hours.  

 

Joe, here's the results of the grep command:

 

 

root@unRAID:~# /root/mdcmd status | strings | grep '^md'

mdVersion=0.95.4

mdState=STARTED

mdNumProtected=9

mdNumDisabled=0

mdDisabledDisk=0

mdNumInvalid=0

mdInvalidDisk=0

mdNumMissing=0

mdMissingDisk=0

mdNumNew=0

mdResync=1953514552

mdResyncPos=110034764

mdResyncPrcnt=5.6

mdResyncFinish=596.6

mdResyncSpeed=51490

root@unRAID:~#

 

 

So it looks OK to these uneducated eyes, the same as Weebotech's plus the parity check stats, right? 

 

I guess at this point I'll let it do it's thing and see what the parity check shows.  If there are no errors and the problem goes away, I guess I won't be able to figure out what happened.  I guess time will tell on this one.

Nice to see the website back up!

 

So, the parity check came back fine, no errors .

 

But I just tried starting up iPhoto and it's telling me my iPhoto library, which is on the server, is corrupt and to restore it from a backup. At the moment, my backup is the parity drive. If the iphoto library is corrupt on the server disk, I'd imagine it's also corrupt on the parity, no?  I'm looking through the iphoto folder itself and the images seem intact, so it must be the database or something else.

 

This brings up a bigger issue as I think about it. If I understand it correctly, using a parity drive is a good safety net if a drive in the server goes bad, but it doesn't do anything for file corruption, does it?  I mean if a file gets corrupted let's say, like the iphoto library, If I swapped out that drive and replaced it with a clean drive, the system would rebuild the disk with the corrupted file I think.

 

One of the reasons I decided to go the unRAID route was to have an automatic and cheap backup system all centralized in one location (I know one should keep a backup offsite...).  This is probably obvious to everyone else, but I'm just now realizing I'll still need to backup vital files. 

 

Come to think of it, wouldn't a RAID 1 mirror system have the same problem with corrupted files, i.e. if the file becomes corrupted, won't it also be corrupt on the mirror? What do you guys do to protect yourselves? I have about a 1TB itunes library and 500mb iPhoto library. I'm thinking these need to be backed up to an external drive from time to time.  Or am I misunderstanding?

Come to think of it, wouldn't a RAID 1 mirror system have the same problem with corrupted files, i.e. if the file becomes corrupted, won't it also be corrupt on the mirror? What do you guys do to protect yourselves? I have about a 1TB itunes library and 500mb iPhoto library. I'm thinking these need to be backed up to an external drive from time to time.

Raid is NOT a backup.  It is a way to deal with a failed disk drive.

 

If you accidentally overwrite a critical file, it is gone, even with mirrored drives and/or RAID.

 

If you value your photos, you'll purchase an inexpensive USB drive and copy the files to it, and keep it off-site from your server.  Or, copy them to writable CD or DVD, and store them off-site.  Otherwise, all it takes is a single fire/flood/tornado/lightning strike and your server and all on it is toast.

 

Read this for more detail: http://lime-technology.com/forum/index.php?topic=2601.msg21033#msg21033

 

If the parity check came back fine, then anything you would re-construct if you were to have a failed drive would be exactly as it is now.  If your software corrupted a file, it will be corrupted when re-constructed.  Parity lets you re-construct a failed disk, not fix files on it.

 

Joe L.

Thanks Joe, that's what I thought.  It's obvious...