Unraid crashes. Kernal?

[kenshinx34]

Hey All,

I'm a long time Unraid user and have not had this type of error before. When transferring files, at some point during the transfer Windows experiences "an unexpected network error". After receiving this error I could no longer ping Unraid, SSH, connect to shares, or pretty much anything. It was clear that Unraid was still on, but nothing was responding. The only way to get it out of this was to hard power it off. This obviously isn't good and forces a parity check each time. I'm currently running the latest build of Unraid 6.1.7. Docker is not running and the only other plugin I have running is Sleep (other than the stock plugins). The drives are all reporting clean SMART stats, and the pool is 6 drives (10TB total) with about 70% utilization.

 

Here's my system build:

M/B: ASUSTeK Computer INC. - M4A88T-M

CPU: AMD Athlon II X3 450 @ 3200

HVM: Disabled

IOMMU: Disabled

Cache: 384 kB, 1536 kB

Memory: 6144 MB (max. installable capacity 16 GB)

Network: eth0: 1000Mb/s - Full Duplex

Kernel: Linux 4.1.15-unRAID x86_64

OpenSSL: 1.0.1q

 

After digging deeper and tailing syslog during a transfer here's what I found (see the attached two logs). At first I thought it had something to do with AFP so I disabled the share with the apple shares. Still the same issue. The transfers range anywhere from 50GB to 250GB which is big, but has never been a problem in the past. This machine has been incredibly stable (uptimes of 8 months no problem). I haven't made any hardware changes recently either. Any help would be awesome!

 

First.txt

Second.txt

[kenshinx34]

Also the CPU is at max 15% utilized so it is not taxed at all and airflow is great.

[kenshinx34]

Well unraid is just about unusable now. I'm really surprised no one has replied to this post.

[RobJ]

Unfortunately, there's not a lot can be said.  You're getting 'general protection faults', not something we can debug remotely.  You can try going back a few versions, see if anything changes.  You can attach your Diagnostics, but I doubt we'll see anything more.  Can you recall what system changes you may have made, just before it began to do this?  Naturally you'll want to do a thorough Memtest, and file system checks, just in case.  You can try a different network card, which may use a different driver.  Obviously, we're grasping at straws here.

[aptalca]

I'd suggest a mem test.

[kenshinx34]

Totally the memory. I am an idiot. How much should I worry about the files that I copied over? If they copied okay what do you think the chances of corruption are/were?

 

[RobJ]

It's hard to say, but certainly there is a chance of corrupted files.  Rsync was involved, and dd and other transfer-related modules were involved, and they were 'tainted'.

 

If possible, I'd delete and recopy, otherwise I'd find some way to verify their bit-for-bit equivalence to the originals - file compares, MD5's, etc.