Emask error during rebuild

December 25, 200817 yr

I just upgraded a 500gb disk to 1tb, during the rebuild (it's still in progress) there was an emask error.

Dec 25 04:47:00 Tower3 kernel: ata3.00: exception Emask 0x10 SAct 0x1 SErr 0x780100 action 0x2

Dec 25 04:47:00 Tower3 kernel: ata3.00: irq_stat 0x08000000

Dec 25 04:47:00 Tower3 kernel: ata3: SError: { UnrecovData 10B8B Dispar BadCRC Handshk }

Dec 25 04:47:00 Tower3 kernel: ata3.00: cmd 60/00:00:27:b9:f0/04:00:19:00:00/40 tag 0 ncq 524288 in

Dec 25 04:47:00 Tower3 kernel: res 40/00:04:27:b9:f0/00:00:19:00:00/40 Emask 0x10 (ATA bus error)

Dec 25 04:47:00 Tower3 kernel: ata3.00: status: { DRDY }

Full syslog below, can anyone help idendify wich disk/controller is the problem? I assume that with this error the rebuilt drive may have corrupted data, should i install the old 500gb drive and run a new parity check?

Thnaks for any help.

December 25, 200817 yr

This looks like just a momentary communications glitch on the bus, while under heavy traffic. I don't believe you have any problems at all. unRAID was not aware of it, as it is detected and dealt with at a lower level. unRAID requested data and got it, although there was a brief delay this one time, while the link was reset. At the lower level, an ATA bus error with evidence of invalid parameters, occurred, so the ATA subsystem paused and reset the link, and after about 2 seconds, the link came back up at full speed and without any further issues, and the appropriate data was returned to the unRAID system.

I've been doing some searching on similar errors, and seeing comments from the experts that the SATA 'system' is somewhat fragile, especially the early SATA chipsets, and subject to a number of glitches. The exception handler has had extra effort put into it to properly take care of these issues, and restore the channel to normal operation. The whole exception Emask reporting mechanism certainly looks scary, too cryptic for almost everyone, but this is a case where it is trying to report as much technical information as possible, and I would guess that in a year or 2, when this whole SATA world stabilizes more, that the error reporting will also be greatly 'simplified', with only very brief messages, as in most other subsystems.

I really don't like these hard resets, but the experts seem to be working very hard on this, to improve the safety and stability of the whole operation.

Apart from seeing this mentioned in the syslog, I don't think you will see any errors. If you are in anyway unsure, you might try mounting the old drive, and comparing the files. I'm not familiar with file compare tools under Linux though, but others here can help.

December 27, 200817 yr

Author

Thanks for the reply, to be safe i installed the 500gb disk again, did a parity check and then tryed to upgrad the disk again, it was rebuilding during the night, in the morning unraid had crashed, i can't access the server with the web browser or telnet, i was tailing the syslog and this is the all i can see: