Comment 49 for bug 1670634

Revision history for this message
bugproxy (bugproxy) wrote : Comment bridged from LTC Bugzilla

------- Comment From <email address hidden> 2017-09-26 07:41 EDT-------
Ok, just checked kernel 4.4.0-97 ... that looks much better:

root@xxxxx:~# uname -a
Linux mclint 4.4.0-97-generic #120-Ubuntu SMP Tue Sep 19 17:27:01 UTC 2017 s390x s390x s390x GNU/Linux
root@xxxxx:~# systool -v -m scsi_mod
Module = "scsi_mod"

Attributes:
uevent = <store method only>

Parameters:
default_dev_flags = "0"
eh_deadline = "-1"
inq_timeout = "20"
max_luns = "512"
scan = "async"
scsi_logging_level = "0"
use_blk_mq = "N"
root@xxxxx:~# systool -v -m dm_mod
Module = "dm_mod"

Attributes:
uevent = <store method only>

Parameters:
reserved_bio_based_ios= "16"
reserved_rq_based_ios= "256"
stats_current_allocated_bytes= "0"
use_blk_mq = "N"

blk_mq() is now turned off by default and that was our main concern! On top, I also started two "big" pdebuild processes (firefox) that so far had the potential to drive the system right into the hang scenario that was the origin cause to write this ticket. The build did not succeed, but the point is that the system did not run into the typical hang -> so I think you can consider this problem being solved!