[stgt] help tgt segfault

[Tomasz Chmielewski]

Tomasz Chmielewski schrieb:
> FUJITA Tomonori schrieb:
> 
>>>> No more I/O errors from open-iscsi side as well.
>>> I'm running a bit different "test" now and I'm getting host resets on 
>>> open-iscsi side every few minutes:
>>>
>>>   connection3:0: iscsi: detected conn error (1011)
>>> iscsi: host reset succeeded

> Anyway, it is something with open-iscsi, as it's reproducible with IET 
> as well, so I'll ask at open-iscsi list.

The solution for such connection errors and host resets is increasing the 
timeout in /sys/devices/platform/host*/session*/target*:0:0/*:0:0:1/timeout  
 
It can be done automatically by a simple udev rule, i.e. to increase it to 720 seconds:

SUBSYSTEM=="scsi", KERNEL=="[0-9]*:[0-9]*", ACTION=="add", ATTR{type}=="0|7|14", ATTR{timeout}="720"


By default, this timeout is 30 or 60 seconds, depending on your distribution.


I observed such resets under the following conditions:
- using DRBD, which in certain circumstances (i.e., just 2 Mbit uplink) is 
  similar to a slow connection between the target and the initiator
- very loaded target (and possibly initiator), even when connected with gigabit links

These resets can ultimately lead to I/O errors and data corruption (although
I had no single data corruption when the target and the initiator were connected 
with a gigabit link, even when host resets were occuring).


See also "connection, host resets, I/O errors eventually (DRBD, but not only)"
thread on open-iscsi list for more info.


-- 
Tomasz Chmielewski
http://wpkg.org
--
To unsubscribe from this list: send the line "unsubscribe stgt" in
the body of a message to majordomo at vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html