I lost power, again. This happens a lot in this area, and I am too stupid to figure out how to set up the ability to shutdown safely from the power button when my UPS kicks in for 5 minutes.

 

When the system came back up, I'm seeing this when checking the status:

 

 

Disk 5 is showing "unformatted" all of a sudden.

 

I'm letting the parity check run (but shit, it takes a long time).

 

What else can I do to check/repair? 

Wait. Be patient. The only action I can recommend taking right now it looking at your syslog, but I'll bet what's happening is that the drive is replaying the transaction log, and will get mounted properly when that is done.

 

The indication of unformatted is WRONG. It just means the drive hasn't yet mounted properly, it will probably fix itself if you wait.

Giving it some time to mount is good advice.

 

If that does not fix the issue, then it might mean there is some file system corruption.  To see if this is in fact the case you can try putting the array into maintenance mode and running

reiserfsck --check /dev/md??

where ?? corresponds to the disk number.  The output of running the command will prove useful in diagnosing such an issue.

Ok. Thanks guys. I'll give it some time.

Right now, I'm missing a TONNE of information. The array should have close to 11TB of data and right now it's only reading at 3TB. I hope this can be corrected.

Right now, I'm missing a TONNE of information. The array should have close to 11TB of data and right now it's only reading at 3TB. I hope this can be corrected.

You might want to carry out the reiserfsck check mentioned earlier on all of your drives.    It seems suspicious that most of them are showing 0 free space.

Thanks, everybody. The reiserfsck check recommended seems to have corrected the problem. I'm not sure what that command actually did, but my data is all back, parity check completed and everything looks good.

 

On a related note, I have the unit hooked up to a UPS giving me about 10 minutes before the unit shuts down. The trouble is it's running headless and I have no way to execute commands when I've lost power (part of the network between other computers and that one lose power in an outage without backup).

 

Is there any way to execute a safe stopping of the array and shutdown from the power button on the tower?

Thanks, everybody. The reiserfsck check recommended seems to have corrected the problem. I'm not sure what that command actually did, but my data is all back, parity check completed and everything looks good.

Interesting - running reiserfsck with the --check option should not actually change anything - merely validate the file system and recommend next actions for any issues found.  What it does do is ensure that any transaction logs are applied so maybe that fixed it.  I guess all is good if it ends well

 

On a related note, I have the unit hooked up to a UPS giving me about 10 minutes before the unit shuts down. The trouble is it's running headless and I have no way to execute commands when I've lost power (part of the network between other computers and that one lose power in an outage without backup).

 

Is there any way to execute a safe stopping of the array and shutdown from the power button on the tower?

There is the powerdown script that can be installed via unMenu.

 

Built-in UPS support has been promised as one of the new features for v6.  Given the announced timescales for v6 release I would expect it to appear in a beta release imminently.  Whether that will also include safe close down via the power button I have no idea but it would seem likely.  You may prefer to wait for that rather than try to fix things on the current setup?

Groovy. I'll keep my eye open for changes in 6 and then make a decision.

 

Thanks again.

There is also the apcupsd plugin for version 5 and for version 6 that also installs the powerdown package.