Re: Large number of abort_task_set followed by conn_close

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Mon, 25 Jun 2012 10:43:36 -0500
Brian Schaefer <brian.schaefer@xxxxxxxxx> wrote:

> Team,
> 
> I'm trying to stabilize my homebrew iSCSI server that I'm using as a
> ESX datastore. Under heavy usage, i see the following behavior:
> 
> Jun 25 10:00:17 fileserver tgtd: abort_task_set(1150) found 177db9 0
> [message repeats ~100 times, increasing the found hex address]
> Jun 25 10:00:20 fileserver tgtd: conn_close(101) connection closed, 0x8cdbe8 70
> Jun 25 10:00:20 fileserver tgtd: conn_close(107) sesson 0x8cdee0 1
> Jun 25 10:00:20 fileserver tgtd: conn_close(101) connection closed, 0x8cc1f8 5
> Jun 25 10:00:20 fileserver tgtd: conn_close(107) sesson 0x8cd400 1

Sounds like the performance of your device is not enough.


> I think this may be due to the fact that i'm using a /dev/loop0
> devices as the backing store:

Maybe. Loopback devices degrades the performance.

> default-driver iscsi
> <target iqn.2012-04.com.example:vmds200>
>         backing-store /dev/loop0
>         initiator-address 192.168.3.2
>         initiator-address 192.168.3.6
> </target>
> 
> In case it's relevant, /dev/loop0 gets created at boot (manually,
> since this was supposed to be temporary anyway). It's a file i created
> with DD, located on an MDADM RAID1 of 2x 2TB disks (2TB total
> storage). The file itself if 500GB.

Why not simply use a file for tgtd?


> I'm running CentOS6 with kernel 2.6.32-220.17.1.el6.x86_64, and
> scsi-target-utils-1.0.14-4.el6.x86_64.
> 
> In order to stabilize the system, would it be better to
> 
> a) migrate to the OSD storage type
> b) migrate to real physical disks (probably another MDADM RAID1)
> c) The loop device isn't the issue, I have something else messed up.
> -- 
> 
> 
> --Brian
> --
> To unsubscribe from this list: send the line "unsubscribe stgt" in
> the body of a message to majordomo@xxxxxxxxxxxxxxx
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe stgt" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html


[Index of Archives]     [Linux SCSI]     [Linux RAID]     [Linux Clusters]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]

  Powered by Linux