Recently, I had a disk fail. I replaced it (it's fine, thank you) and at some point after the data rebuild disk 8 (serial number ending in 9554) showed as unformatted. I followed steps outlined in the forums for reiserfsck --check and reiserfsck --rebuild-tree on the disk using /dev/mdX. Once it completed, there was a folder called lost+found on the disk. I still have that disk and its contents. Unfortunately, I don't recall what happened after that. I apologize for not keeping better notes.

 

At some point after the rebuild, I know the disk showed errors and was disabled. I replaced the SATA cable for no reason other than I thought it would help. Last night I replaced the failed disk with a new pre-cleared disk, and started the data rebuild. I went to bed and woke up to yet another failed disk with 384 errors.

 

I'm using Dynamix. The Utils > System Log is empty. Clicking Log in the upper right displays tail -n 40 -f /var/log/syslog. I get some kind of database error when searching for forum topics on that. I've added the most recent log from /logs, plus the oldest. They are essentially the same configuration.

 

Where to go from here?

 

v 5.0.4

syslog-20150909-192109.zip

syslog-20150916-192528.zip

To see if the issue is with the physical slot in the middle 5 in 3 cage, I moved the brand new red balled disk to a different slot. It's the disk ending in 7YRK. I moved into second from the left, bottom row. Also a Supermicro cage, but a different one.

 

It's still red balled, but correctly labeled as disk 8 and sdl.

The only ways to clear a redball is rebuild or new config.

The only ways to clear a redball is rebuild or new config.

 

The disk won't spin up at this point. I'd be glad to start a new rebuild.

The only ways to clear a redball is rebuild or new config.

 

The disk won't spin up at this point. I'd be glad to start a new rebuild.

What do you mean by that?  If the disk won't spin up then its obviously toast and needs to be replaced.

After removing the disk from the server, I was able to read the disk contents with DiskInternals Linux Reader. So strange.

I've been using the new bjp preclear script. Not even sure how that could be a problem. I'm going to try the old one.

You're bouncing all over the place here... 

 

Disk errors during operation and/or parity rebuilds are caused by hardware issues, and certainly is not caused by the preclear script.

 

Preclear isn't strictly required to rebuild a drive (only to test to make sure that it works correctly, and is error free)

 

 

Ah, fuck it.