Well, the port multiplier replacement came in and all installed just fine. Â Good news is that it was the PMP that had failed. Â Bad news is that when I got everything configured as it was before the failure, unRAID decided that those three drives were all new and started clearing them!
I hit the reset button on the unRAID box after only a few seconds…it never budged past 0%, but damage was done. Â All drives subsequently showed back up as unformatted when unRAID resumed operation. Â I had lost roughly 750GB of TV episodes and my entire NAS drive.
Fortunately, I started a recovery process that, so far, appears to be promising. Â I performed it first on the NAS drive. Â It involves using reiserfsck commands that are not for the faint-hearted. Â However, considering my predicament, I had no choice.
I started out with the following commands that basically just rebuilds the resierfs directory structure. Â That’s probably dumbing it down a lot since I’m about the farthest thing from a linux guru. Â
root@MEDIASVR:/# /root/samba stop
root@MEDIASVR:/#Â reiserfsck –scan-whole-partition –rebuild-tree /dev/md4
unRAID has to have the array started, but samba sharing has to be stopped in order for it to begin the rebuild. Â After about 5 hours, I had a single lost+found folder that contained many more folders named with seemingly random numbers and over 100 randomly numbered files with no file extension. Â After perusing a few of those folders, it appeared as all of my old directories were in them. Â I have no idea what the no-extension files are, however. Â I’m hoping they’re just old deleted files that were recovered after the fact.
I was a bit more at ease until trying the next drive. Â This was the one that contained the bulk of my TV episodes and after starting the rebuild, it ended shortly after upon hitting a bad block. Â I now had to do aÂ
root@MEDIASVR:/# /sbin/badblocks -b 4096 -o badblocks.lst /dev/sde
to generate a list of all the bad blocks that the drive could discover. Â This lasted about 5 hours – the same length as a recover. Â Once finished, I had to make reiserfsck aware of this list and modify my rebuild command like so,
root@MEDIASVR:/#Â reiserfsck –scan-whole-partition –rebuild-tree -B badblocks.lst /dev/md5
This allows the bad blocks to be skipped or reallocated – I’m not really sure…I just know that reiserfsck knows when and what to do at bad block. Â As I write this I have about 35 minutes to go on the rebuild before I’ll see what the damage is. Â Considering my TV show folders begin at three levels deep, I think I should be in pretty good shape at retaining the directory structure I had. Â I’ll be doing an RMA on this drive once I get everything moved. Â Seagates have 5-year warranties and this one is not even two yet.
After spending a few minutes checking that drive when it’s done, I’ll immediately begin with the third drive. Â It was, at most, about 2/3 capacity remaining and was a spillover for the other TV episode drive. Â I can only hope for no more bad blocks there so I can get this done tonight and spend tomorrow getting it all back in order to use, while finding out what I’m missing. Â I’ll be sure to update tomorrow.
In other news, I received a P3 Kill-A-WATT P4400 device that I had ordered yesterday. Â The Kill-A-WATT plugs into a standard 3-prong electrical outlet and allows you to monitor all the characteristics of what is coming out of that outlet to the device that plugs into the Kill-A-WATT. Â I can get KWh, line condition, voltage, etc. Â I plan on using this on the most used electrical devices in the house to see where we can cut some corners. Â Computers, applicances, stereo equipment, etc. will eventually all be tested. Â I may start a small index so that others can compare to it and see what their devices are drawing while on. Â More on that later.