Can not connect raidz2

[LordNicky]

Good day!

Problem on FreeBSD 9.0. Error begins with the string

Fatal trap 12: page fault while in kernel mode

Full dump of the error here:
https://hotfile.com/dl/178923711/316b1b9/core.txt.2.html

System works 6 months, updated regularly, the last time at the weekend. Reboot is successful and 4 days worked flawlessly. Smart at all HDD under smartd before an error occurs was OK. The system is on a separate disk. On the 4 hard drives assembled raidz2. At first I assumed the error on the system disk, and run in single mode fsck -y. The same result, but if you take away from rc.conf zfs_enable=YES, then fine loaded and an error when you try to run any zpool, except help.

I tried to connect to another sata controller - the same error, I tried to pull out all the HDD's, zpool starts and shows that he does not see the HDD's... The same result with a single hard disk drive connected. I scanned the memory with memtest, using different versions and CD's. Reloaded at different stages of the test. Once passed through. Error does not show. Replaced the memory. The same situation. Now scanned with Goldmemory. Goldmemory scanned completely and without errors. During the 4-hour scan never flew. Next replace the power supply. The same thing. I have also tried to rearrange these drives to another computer. The system booted and also crashed. Also tried to boot from the CD FreeBSD 9.0. When forced import array same error. Now try to boot from the CD with other architectures.

The main question - how to get the information from the raid or fix it? Where to dig further?

I recently started working with FreeBSD, so please forgive that I do not know what some important things. I would be grateful for links to similar problems. Thanks in advance!

 

[Terri_Kennedy]

I recently started working with FreeBSD, so please forgive that I do not know what some important things. I would be grateful for links to similar problems. Thanks in advance!

First, for posting text files like this, I suggest a site like pastebin.com. The more steps someone has to take to view your text, the less likely they are to help.

You're running ZFS on an amd64 kernel with only 2GB of real memory. Many people will tell you that ZFS likes lots of memory. Others will tell you that you can get by with less memory if you carefully configure some ZFS tuning options.

It looks like the trigger for this panic was:

(ada1:ata3:0:0:0): Synchronize cache failed
(ada2:ata4:0:0:0): Synchronize cache failed

Since this seems to have happened on two drives at the same time, I'd suspect the disk controller or cable (if a multilane cable), or a power supply problem.

I'm not familiar enough with ZFS internals to see what caused the trap 12 panic. In general, I'd hope to see a panic such as "assertion failed", which would at least indicate that the kernel knew what was wrong. A trap 12 means that the kernel "went off into the weeds" and tried to execute code that simply wasn't there.

It looks like you're running an unmodified 9.0-RELEASE kernel. You might try downloading the 9.1-RC3 disc image from here to see if you're hitting a bug that has already been fixed.

If you're still seeing the problem in 9.1-RC3, I'd suggest asking on freebsd-fs@[url] as you're more likely to reach a developer there.