One bad SSD in a pool- two failures!

[dtempleton]

I came to posses three small Toshiba SSDs (256g) when I dismantled my hackintosh in favor of an upgraded board and CPU for unraid (that went smoothly!)

 

I decided to put them into new pools that would separate vms and appdata, and use two of them for a pool for the traditional cache buffer.  One of these drives, the old UNRAID cache, had begun throwing CRC errors (about one a week, it seems, and now 50) so I thought that being in a pool with a sounder identical SSD would be a safe thing.  I mean, you have two redundant drives and you lose one you're still OK, right?

 

So I started having problems with rsync for files on this pool ('read only file system' errors) and the log was full of btrfs errors.  Uh Oh.  On reboot, both the SSDs in the pool are unmountable with no file system.  All the data (not much, mostly testing) was just gone.

 

Reformatting the SSDs using default btrfs remounted them and re-formed the pool.  SMART reports show one is still fine and the other has a few more CRC errors.

 

I can't see what happened.  If one SSD poops out, how did that cause the other one to lose its FS?  I'm thinking that with an iffy SSD I'm better removing it that trying to use it to create some redundancy.

 

Any thoughts?

 

Thanks!

[JorgeB]

8 hours ago, dtempleton said:

I can't see what happened. 

Neither can we without the diagnostics from when it happened, as for the CRC errors it's usually a bad SATA cable.