process kills and hangouts

[emi]

i have change the os from debian to freenas to use zfs sensible. Now is the system complete and all services run! The system is only for backup services! I have rsync tasks, nfs+samba shares and one iscsi lun on one vol. All services run on the same zfs pool. i use a lot of memory and the cpu load, caching is ok. The system is up2date(FreeNAS-11.0-U3 (c5dcf4416)) and have two Xeon E5506 cpus. The controller is a LSI MegaRAID SAS 9261-8i anbd all disks are raid0 arrays.

Code:

root@XXXXXXX:/var/log # mfiutil show volumes
mfi0 Volumes:
  Id  Size  Level  Stripe  State  Cache  Name
mfid0 (  930G) RAID-0  64K OPTIMAL Writes
mfid1 (  930G) RAID-0  64K OPTIMAL Writes
mfid2 ( 1862G) RAID-0  64K OPTIMAL Writes
mfid3 (  930G) RAID-0  64K OPTIMAL Writes
mfid4 (  930G) RAID-0  64K OPTIMAL Writes
mfid5 (  930G) RAID-0  64K OPTIMAL Writes
mfid6 (  930G) RAID-0  64K OPTIMAL Writes
mfid7 (  930G) RAID-0  64K OPTIMAL Writes
mfid8 (  930G) RAID-0  64K OPTIMAL Writes
mfid9 (  930G) RAID-0  64K OPTIMAL Writes
mfid10 (  930G) RAID-0  64K OPTIMAL Writes
mfid11 ( 1862G) RAID-0  64K OPTIMAL Writes
mfid12 ( 1862G) RAID-0  64K OPTIMAL Writes
mfid13 ( 1862G) RAID-0  64K OPTIMAL Writes
mfid14 ( 1862G) RAID-0  64K OPTIMAL Writes
mfid15 ( 1862G) RAID-0  64K OPTIMAL Writes

But i have hangouts and process kills. VMs get to read only file systems and some process (which using nfs and samba shares) have timeouts.

Code:

(0:3:0/0): READ(10). CDB: 28 00 00 00 ae 58 00 00 01 00
(0:3:0/0): Tag: 0xa95947, type 1
(0:3:0/0): ctl_process_done: 179 seconds
(0:3:0/0): READ(10). CDB: 28 00 00 00 ae 58 00 00 01 00
(0:3:0/0): Tag: 0xa9594a, type 1
(0:3:0/0): ctl_process_done: 139 seconds
(0:3:0/0): READ(10). CDB: 28 00 00 00 ae 58 00 00 01 00
(0:3:0/0): Tag: 0xa9594d, type 1
(0:3:0/0): ctl_process_done: 99 seconds
ctl_datamove: tag 0xa95994 on (0:3:0) aborted
ctl_datamove: tag 0x385b3604 on (3:3:0) aborted
ctl_datamove: tag 0x385b3605 on (3:3:0) aborted
ctl_datamove: tag 0x385b3606 on (3:3:0) aborted
ctl_datamove: tag 0x385b3607 on (3:3:0) aborted
ctl_datamove: tag 0x385b3608 on (3:3:0) aborted
ctl_datamove: tag 0x385b3609 on (3:3:0) aborted

What can i do? What is wrong?

 

[Artion]

LSI MegaRAID SAS 9261-8i anbd all disks are raid0 arrays

I don't know if this can help but I see in the forum that for LSI raid controllers is recomended to use them in IT mode that requires flashing them with an appropriate firmware.

 

[Ericloewe]

The SAS 926x constrollers do not support IT mode.

 

[Artion]

Just for curiosity: why you have a pool with two raidz1 vdevs which one have 10 drives and the other 6? Is there any perfomance gain by having this configuration?

 

[emi]

Thanks for the answers!! #artion: The disks have different sizes. i had try it with one raidz1 and so but the problems were the same. The disks have different sizes.
my last test was wit an other controller (LSI SAS 9205-8e in it mode) . but i think the controller had problems with the expander. :-(