I upgraded my 4.7 server to 5.0 last week. I didn't do any of the beta since I felt I couldn't afford to lose anything on my main server. The upgrade went smooth. Everything came up in the right place and I have been running fine since the upgrade until I tried to add a new user share last night. When I went to add a new share, the server crashed with a kernel panic. I forced a reboot, everything came back online just fine and thinking it may have been a fluke, I tried adding a new share again. Same thing, kernel panic. I never had any issues with 4.7. I was able to add and modify my shares to my hearts content.

 

I know I have nothing to provide to get help just yet, so my first question is, how can I get a log to see what may be happening? Since the syslog doesn't' persist after a reboot, and I can't access the server after the kernel panic, what should I do?

 

My build:

 

Intel BOXDG43NB Intel G43 ATX

Intel Pentium Dual-Core E6500 Wolfdale 2.93GHz

CORSAIR CMPSU-400CX 400W

4GB RAM

See SYSTEM LOG in my sig. Save a copy of the entire syslog using the unRAID GUI under Utils and then click the Log button on the top right of the unRAID GUI the open a window that will display and capture the ongoing syslog in real time. Leave this window open and  copy and paste it's contents into a text file after a crash. You will have 2 files that show the log from boot to crash. 

Finally got around to testing things out. It seems to kernel panic with a couple of things. Creating a new share is one and panics when it tries to create the directory at /mnt/user. It also panics when I try and change the settings on a share. Attached is a log from boot to crash. The cause of the crash was enabling AFP for a share.

panic.txt

Disable all addons and/or start in safe mode.

I don't have any addons installed at the moment. Also, safe mode? I didn't know unraid had a safe mode.

Safe mode is new.  You'll need to have a monitor and keyboard hooked up to access it

 

Sent from a mobile device, sorry for any typos.

 

 

I booted in safe mode and am still getting a panic. Attached is the log from this time. Again, turning on AFP for a user share and hitting apply.

panic2.txt

Still can't figure out what is going on, but I'm starting to think a filesystem issue or something. This is only happening on a few shares and any new shares I try and create that involve all disks. I can modify and do whatever I want to certain shares that are limited to a single disk or 2. I figured this out when I tried to copy some files to a certain share. The copy itself caused a kernel panic. I then tried to mv via telnet some files from one disk to another and got a kernel panic. I ran smart tests on the disks and everything checks out. I'm currently running reiserfsck --check on them as well. See the link below for a couple of pics I took of 2 of the panics. One was caused by enabling AFP on a share that includes all disks. The other is when I tried to move files from one disk to another. I suspect filesystem issues with one of these 2 disks. Thoughts??

 

http://www.flickr.com/photos/icedtrip/sets/72157636526537334/

I don't have any addons installed at the moment. Also, safe mode? I didn't know unraid had a safe mode.

 

Oct 12 17:49:37 Tower logger:  file /boot/packages/dmidecode-2.10-i486-1.txz: already exists
Oct 12 17:49:37 Tower logger:   upgradepkg --install-new /boot/packages/dmidecode-2.10-i486-1.txz ... 
Oct 12 17:49:37 Tower logger: +==============================================================================
Oct 12 17:49:37 Tower logger: | Installing new package /boot/packages/dmidecode-2.10-i486-1.txz
Oct 12 17:49:37 Tower logger: +==============================================================================
Oct 12 17:49:37 Tower logger: 
Oct 12 17:49:37 Tower logger: Verifying package dmidecode-2.10-i486-1.txz.
Oct 12 17:49:37 Tower logger: Installing package dmidecode-2.10-i486-1.txz:
Oct 12 17:49:37 Tower logger: PACKAGE DESCRIPTION:
Oct 12 17:49:37 Tower logger: # dmidecode (DMI table decoder)
Oct 12 17:49:37 Tower logger: #
Oct 12 17:49:37 Tower logger: # dmidecode is a tool for dumping a computer's DMI table (some say
Oct 12 17:49:37 Tower logger: # SMBIOS) contents in a human-readable format.  This table contains a
Oct 12 17:49:37 Tower logger: # description of the system's hardware components, as well as other
Oct 12 17:49:37 Tower logger: # useful pieces of information such as serial numbers and BIOS
Oct 12 17:49:37 Tower logger: # revision.
Oct 12 17:49:37 Tower logger: #
Oct 12 17:49:37 Tower logger: # dmidecode was written by Alan Cox and Jean Delvare.
Oct 12 17:49:37 Tower logger: #
Oct 12 17:49:37 Tower logger: Package dmidecode-2.10-i486-1.txz installed.
Oct 12 17:49:37 Tower logger: 
Oct 12 17:49:37 Tower logger: 
Oct 12 17:49:37 Tower logger: success
Oct 12 17:49:37 Tower logger:  file /boot/plugins/webGui-latest.txz: already exists
Oct 12 17:49:37 Tower logger:  file /tmp/plugin-tmp: successfully wrote INLINE file contents
Oct 12 17:49:37 Tower logger:   /bin/bash /tmp/plugin-tmp ... Verifying package webGui-latest.txz.
Oct 12 17:49:37 Tower logger: Installing package webGui-latest.txz:
Oct 12 17:49:37 Tower logger: PACKAGE DESCRIPTION:
Oct 12 17:49:37 Tower logger: Package webGui-latest.txz installed.
Oct 12 17:49:37 Tower logger: 
Oct 12 17:49:37 Tower logger: success
Oct 12 17:49:37 Tower logger:  plugin successfully installed
Oct 12 17:49:37 Tower logger: Installing user plugins

 

 

This package is not part of the unRAID distribution. Prepare a clean install on a formatted flash.

This package is not part of the unRAID distribution. Prepare a clean install on a formatted flash.

 

I guess I wasn't thinking of the new webgui as being an add-on. Also, I don't recall installing dmidecode. Was this installed along with the new webgui?

 

Either way, I started from scratch. I backed up my flash drive. Oddly, when I did, it wouldn't let me copy over one of the share config files due to corruption. Well, I formatted the flash drive, did a fresh install, but this time left off the share configs and reconfigured everything. Everything came up fine, but I am still having a panic when I interact with certain shares...the 2 shares that use all of the installed drives. Earlier, I mentioned that it paniced when I tried to move files directly from 1 disk to another. Again, the smart reports are fine with these 2 drives and the "reiserfsck --check" for both drives were fine too.

 

Attached is the most recent syslog up until crash.

panic3.txt

I have narrowed it down to a single disk. It's odd, but the panic occurs whenever a new directory is made on the top level. I can make folders and copy files to all other levels of the drive. Both the smart report and reiserfsck --check both came out fine for this disk.

See here: http://lime-technology.com/forum/index.php?topic=28484.0

See here: http://lime-technology.com/forum/index.php?topic=28484.0

 

Thanks. I ran through all of my top level directories on all of my drives and found that several had the extended attributes set on the directories. I ran Tom's script and cleared them all up (just triple checked to make sure) and I'm still getting the kernel panic trying to enable AFP or creating a directory on the top level of this particular drive.