Error in logs with 3ware RAID, what could be wrong?

Question

RAID

Hello.
I'm scratching my head and can't find any information on the problem, so I decided to ask a question here.
There is a server that works as a "SAMBA-ball" on CentOS, an error periodically starts to appear in the logs several times per second:

Jan 21 08:14:19 localhost kernel: Descriptor sense data with sense descriptors (in hex):
Jan 21 08:14:19 localhost kernel:         7f 00 00 00 00 00 00 28 00 00 00 00 00 00 00 00 
Jan 21 08:14:19 localhost kernel:         00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 
Jan 21 08:14:19 localhost kernel:         00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 
Jan 21 08:14:23 localhost kernel: 3w-9xxx: scsi4: ERROR: (0x03:0x101A): Retry queued command:.
Jan 21 08:14:23 localhost kernel: sdd: <<DEFERRED>> [descriptor]: sense key: No Sense
Jan 21 08:14:23 localhost kernel:     Add. Sense: No additional sense information

If you restart the server, then it spits out an error on the 3ware account:

Jan 21 09:36:10 localhost kernel: 3w-9xxx: scsi4: AEN: ERROR (0x04:0x002A): Verify failed:unit=0.
Jan 21 09:36:10 localhost kernel: 3w-9xxx: scsi4: AEN: INFO (0x04:0x003D): Verify paused:unit=0.

Today, SAMBA generally hung up and it was not possible to reboot it, it also did not work to kill the process, only a hard reboot helped.
At the same time, disks through 3ware are normally detected in the system.

If you run the tw_cli show command, then an error occurs:

[[email protected] ~]# tw_cli show

(0x0B:0x000A): Invalid drive ID


Ctl   Model        Ports   Drives   Units   NotOpt   RRate   VRate   BBU
------------------------------------------------------------------------
Ошибка сегментирования

I understand that the problem is not in the correct operation of the 3ware RAID controller, but I cannot understand what can be done. If there are specialists in this matter, I will be grateful for your help.

Answer 1

Item #0: Do you have integrity-checked backups of server content? Without this, I would not have experimented with the system, especially in terms of operations involving recording.
You have an interesting output in terms of what is displayed Retry queued command:.without displaying the command. In theory, there should be something like 0x51 .
Judging by the dump, the problem occurs with the read operation (read) - 28 M OM MM READ(10).
It would be good to learn what your 3ware manual says about this.
In addition, if you have backups and suddenly have the opportunity to boot Windows (for example, from a USB flash drive), then run a program like R.tester and see what it says about the state of the disks.
First of all I would check slot 27 (or 26). Unless, of course, you have such a number of disks at all and my assumption that this is numbering is correct.