I'm running unraid pro 4.7, with 20 disks, and am having a few issues.  It started off with a legitimate drive failure (well it was starting to fail), disk12, so I replaced it, rebuilt the data, ran parity check, and everything was fine (I've done this a few times in the past, both for drive failures and for upgrades, and have never had an issue).  I shut the server to move its physical location (about 8 feet), and ever since then I turned it back on I've been having issues.

 

Initially some disks weren't detected, then disk12 (the one that was just rebuilt) was labeled as uninstalled, and then the web interface would sometimes not load (even if I tried to force it to start in command line), which then had me doing improper shutdowns.

 

To try to narrow the problem down I initially unplugged all of the drives, started the server (checked to see if I could get to the web interface even though no drives were attached: I could), then powered it down, and then started slowly adding back all of the drives to the server (3-4 each time, making sure that they were detected in the web interface). 

 

It eventually did detect all of them, except for disk12 (wasn't labeled as missing, but i think it said not installed... basically it was thinking the data hadn't been rebuilt on it, even though it had), but the array was working fine since it was just re-creating the missing data by using parity.  After I pulled off some data that I needed right away, I shut the server down, waited a few minutes (checked drive, cables, etc.) and powered it back on, and since then it's been having trouble re-detecting all of the disks again; my idea of slowly re-adding them all back to the array has not worked this time.

 

I notice there are certain problem disks (12, 13, 16, 17, 18) , and I noticed that these problem disks are drives that have been replaced over time (sometimes I'll see an another serial# below the disk when it's not detected, oh and sometimes it detects the 750gb drive as a 2.147TB drive). 

 

The RAID controllers I use detect them all of the drives every time during bootup in bios (always says bios installed successfully), and all of the connections seem tight and solid.  I ran a memtest for around a 24 hour period and it did I think 7 complete passes and detected no errors.  I also went into the go file and commented out everything except for /usr/local/sbin/emhttp &

 

At this point I actually think all of the data is on all of the drives (even drive 12 which had been rebuilt successfully, and I was pulling data off of it at one time, only for it to say not installed after a reboot), but I just don't have anywhere to put it all if I want to rebuild the server from scratch again (which probably wouldn't be the worst idea since I have upgraded the server so many times w/more controllers, drives, different drives, new versions of unraid, etc.).

 

I'm thinking some of the improper shutdowns (more than once I just flipped the powerswitch as the powerdown script I had wasn't working and 'shutdown now' would often just hang for minutes, I'd hit ctl+f2 to bring up another session, log back in, and would try to shut it down again, only for it to hang again) screwed things up, but I have no idea how to fix them.

 

I've been using Unraid for a few years and haven't ever had an issue like this one before; it's not the end of the world or anything, but I am wondering what's going on, and if this is even fixable.

 

I attached a syslog from last night (syslog-1) that shows disk12 removed (even though it's not), and disk17 and 18 as missing; and then I just turned on the server again and you can see how it's a little different (disk12 is now being seen, but as 2.147TB even though it's only 750GB, and disks 17 and 18 are now being seen in the array, but disk 16 is now missing; it also seems to think that disk11, 12, and 13, all have the same serial #).

 

It's all so random, I'm confused.  Any help?

 

Nate

syslogs.zip

What power supply are you using?  (specific make and model)

 

With 20 drives, is it possible you are at its limit of capacity?  (Especially if it is not a high amperage single-rail 12 volt supply)

 

Have you performed a memory check recently? Yes, I see you did.

 

Since this all started when you moved the server, have you tried to re-seat the disk-controller cards?

 

Joe L.

 

Power Supply: Corsair TX750W http://www.corsair.com/tx750w.html

 

As for the limits of capacity, I suppose it's possible, but I've had 20 drives since a week or two after Unraid upgraded to that capability, and have never had an issue.  In my last drive upgrade (2 years ago?) I swapped out four 750gb drives for four 2TB, the 2TB models being the WD green models that consume less power-so if anything a power issue should have happened before that upgrade.

 

I realize power supplies go bad though, and the power supply tester I have I don't completely trust (it's a few years old and was only $20), but if I only plug in drives 12-20 (8 drives should definitely not be an issue), unraid is still not seeing them all-so it sees the first 12 drives (parity thru 11) every time no matter what else is plugged in, but if only the last 8 are plugged in, it never sees all 8.

 

Okay, I just found the common denominator, disks 12, 13, 16, 17, 18, 19, are all plugged into SuperMicro SAT2-MV8 (disks 14 and 15 are plugged into one of my two Adaptec 1430SA controllers-I thought I had installed everything in sequential order, I guess not). 

 

I just re-seated the PCI card for SuperMicro SAT2-MV8, like you suggested, and Unraid is now seeing all 20 drives again (disk12 is red and needs to be rebuilt again for some reason, but that's no biggie), wow, this is great. 

 

I'm not going to lie though, I'm nervous to shut it off for fear that it won't work again.  The card was screwed into the slot and seemed fairly tight, and upon booting the controller card detected all 6 drives; reseating it must have done something though-strange, I don't understand that at all.  To me if the card wasn't set properly, the computer either shouldn't detect it, and the card certainly shouldn't be detecting all the drives that are plugged in.

 

If you don't mind, I'm not going to mark this as solved just yet, I'm going to rebuild the drive, do a parity check, move some data around and off the server, and then shut it down later this weekend.  On Monday or Tuesday after I re-start and it sees the drives again I'll mark it as solved.

 

Thanks a lot Joe.

 

Nate