Other CTLD issue not able to terminate iSCSI connection

N

Nutz

Hi All,

I'm a facing an issue with at least one iscsi connection on a freebsd 11.1 with ZFS/ZVOL
When the initiator logout the session the request is well interpretated by Freebsd but i got the message :

Code: 
WARNING: 172.44.105.55 (iqn.1993-08.org.debian:01:e0d55e173e35): waiting for CTL to terminate 12 tasks
WARNING: 172.44.105.87 (iqn.1993-08.org.debian:01:e0d55e172ea7): waiting for CTL to terminate 10 tasks
WARNING: 172.44.106.192 (iqn.1993-08.org.debian:01:e0d55e173a41): waiting for CTL to terminate 10 tasks
WARNING: 172.44.109.208 (iqn.1993-08.org.debian:01:e0d55e1a139f): waiting for CTL to terminate 10 tasks

There are several initiators name as the older connections are not terminated correctly.

From dmesg i have a lot og message like this :
Code: 
(5:99:0/104): READ(10). CDB: 28 00 02 14 e1 82 00 00 01 00
(5:99:0/104): Tag: 0x005a, type 1
(5:99:0/104): ctl_datamove: 97 seconds
ctl_datamove: tag 0x005a on (5:99:0) aborted
(5:99:0/104): Tag: 0x0033, type 1
(5:99:0/104): ctl_datamove: 97 seconds
ctl_datamove: tag 0x0033 on (5:99:0) aborted
(5:99:0/104): READ(10). CDB: 28 00 02 14 dd b3 00 00 39 00
(5:99:0/104): Tag: 0x0045, type 1
(5:99:0/104): ctl_datamove: 97 seconds
ctl_datamove: tag 0x0045 on (5:99:0) aborted
(5:99:0/104): Tag: 0x002e, type 1

So first, whats does it means the number at the beginning ? i did find any clue nor disk or lun..with ctladm devlist, reportlun..etc..

I increased debug level with kern.cam.ctl.debug :

Code: 
Oct 13 16:06:11 <0.2> str-srv kernel: (10:167:0/184): SCSI sense: UNIT ATTENTION asc:29,1 (Power on occurred)
Oct 13 16:06:11 <0.2> str-srv kernel: (11:167:0/184): TEST UNIT READY. CDB: 00 00 00 00 00 00  Tag: 0x6/1
Oct 13 16:06:11 <0.2> str-srv kernel: (11:167:0/184): CTL Status: SCSI Error
Oct 13 16:06:11 <0.2> str-srv kernel: (11:167:0/184): SCSI Status: Check Condition
Oct 13 16:06:11 <0.2> str-srv kernel: (11:167:0/184): SCSI sense: UNIT ATTENTION asc:29,1 (Power on occurred)
Oct 13 16:07:21 <0.2> str-srv kernel: (6:26:0/36): Task Action: I_T Nexus Reset
Oct 13 16:07:21 <0.2> str-srv kernel: (6:26:0/36): CTL Status: Command Completed Successfully

(19:187:0/87): SYNCHRONIZE CACHE(10). CDB: 35 00 00 00 00 00 00 00 00 00  Tag: 0x55/1
(19:187:0/87): CTL Status: SCSI Error
(19:187:0/87): SCSI Status: Check Condition
(19:187:0/87): SCSI sense: UNIT ATTENTION asc:3f,e (Reported LUNs data has changed)

so now, i dont really know where to dig...i try to find out if there is a way to check the queue tasks from ctl to see which one is stuck
So if you have any ideas...

BR,
Nutz!
 
Hi Again,

As ctl is a kernel process i 'm trying to use Dtrace, does anyone know which probe i have to use to dig deeper ?
i tried with fbt:ctl::: ctl_queue & ctl_close...

and if anyone knows how to identify device with this
Code: 
(11:167:0/184)
 
Hi Sirdice,

Yep, i know, and we are working to update our storage server from 11 to 12 but we cannot do this right now on the server where the issue is occured.
 
I can understand that. But you also have to understand you might be hitting a bug that's been fixed in newer versions.
 
You must log in or register to reply here.
Back
Top