Cache Drive Failed. Web Interface doesn't work. [Solved]

[tea_rev]

Well, after some extensive troubleshooting, searching and frustration, I've determined that my cache drive is dead. (Well, not dead, but "frozen". Apparently a common problem for OCZ drives with sandforce controllers.) This wouldn't be too much of a concern, except that it seems to have taken out my web interface in the process, ironically just about when I had planned to add several new drives to the system.

 

So now at this point, I'm stuck with a webgui that looks like this:

 

UnMENU still works great, except that I can't manage my drives from there. File sharing works fine (streaming video from it right now actually) and all the packages that i've checked so far have no issue, the only thing that appears broken is the webgui.

Would it be possible to repair the web interface directory without reinstalling UnRAID? (already attempted the simplefeatures interface, and that's just broken in a different way.)

Has anyone experienced a similar problem who might be able to give me some guidance?

 

Issue happened while running Beta 12a, but I've since upgraded to 13.

 

Syslog attached.

syslog-2011-11-18.txt

[dgaschk]

Shutdown safely:

Three safe ways to power down:

The unRAID webGUI

The powerdown script (available via unMenu). Use powerdown -r to restart.

http://lime-technology.com/wiki/index.php? title=Console#To_cleanly_Stop_the_array_from_the_command_line

 

There are no other safe ways to power down!

 

Remove the cache drive and reboot.

[tea_rev]

Ah if only it were that simple. I probably should have posted what troubleshooting steps I have already attempted.

 

I first noticed the problem when attempting to upgrade my array. There was a red led lit on the back of the cache drive which I had previously discovered (with a separate drive of the same model) indicates that it is locked (and therefore useless). So I powered down the system using SSH and "powerdown -h now". I then booted the system back and everything was still working fine, (including the web interface). So, I just figured no big deal, no cache drive anymore, might as well upgrade my system. shutdown again (always using powerdown -h or the unmenu shutdown button) and popped in another drive. Boot back up, click the refresh button on the web interface and I get the page I posted above.

 

At this point I have attempted all of the following recommendations I could find on this board and elsewhere.

 

Reboot

Removed the new drive and reinstalled the cache drive. Nothing, so I removed the cache drive again.

Checked is there is enough free RAM (the machine has 8GB and after boot almost 7GB are free)

Run Memtest for 26hours

Upgraded/downgraded UnRAID versions.

Updated all packages, including unmenu itself.

Cloned the USB drive, reformatted and cloned the data back.

Installed another drive in place of the dead cache drive.

Swapped SATA cables.

Killing and restarting the web interface with

killall emhttp
nohup /usr/local/sbin/emhttp &

(This does not bring back up the web interface as I simply receive a segmentation fault.)

 

Everything in these¹threads² (even rewriting my go script from scratch in vim)

Used multiple browsers, multiple computers, clearing cache and cookies on all of them.

 

I was planning on attempting the Trust My Array procedure, but it is not possible with no web interface and no CLI equivalents posted.

 

 

Basically everything short of starting over from scratch (not really looking forward to reconfiguring all of my shares and packages)

[tea_rev]

If only I had found THIS post sooner!

 

Turns out this was caused by the php package installed by unmenu. I've since removed the package and the issue is solved.

 

The php package and the cache drive failing were separate and purely coincidence, thus causing diagnostics and troubleshooting to be quite difficult.