[LU-5526] recovery-mds-scale test failover_mds: dd: No space left on device

Related — Wed, 5 Aug 2020 21:48:42 +0000

While running recovery-mds-scale test failover_mds (MDS failed over 1 time), client load on one of the clients failed as follows:

2014-08-16 20:44:59: dd run starting
+ mkdir -p /mnt/lustre/d0.dd-shadow-41vm5.shadow.whamcloud.com
+ /usr/bin/lfs setstripe -c -1 /mnt/lustre/d0.dd-shadow-41vm5.shadow.whamcloud.com
+ cd /mnt/lustre/d0.dd-shadow-41vm5.shadow.whamcloud.com
+ sync
++ /usr/bin/lfs df /mnt/lustre/d0.dd-shadow-41vm5.shadow.whamcloud.com
++ awk '/filesystem summary:/ {print $5}'
+ FREE_SPACE=14195328
+ BLKS=1596974
+ echoerr 'Total free disk space is 14195328, 4k blocks to dd is 1596974'
+ echo 'Total free disk space is 14195328, 4k blocks to dd is 1596974'
Total free disk space is 14195328, 4k blocks to dd is 1596974
+ load_pid=3715
+ wait 3715
+ dd bs=4k count=1596974 status=noxfer if=/dev/zero of=/mnt/lustre/d0.dd-shadow-41vm5.shadow.whamcloud.com/dd-file
dd: writing `/mnt/lustre/d0.dd-shadow-41vm5.shadow.whamcloud.com/dd-file': No space left on device
1213957+0 records in
1213956+0 records out
+ '[' 1 -eq 0 ']'
++ date '+%F %H:%M:%S'
+ echoerr '2014-08-16 20:52:06: dd failed'
+ echo '2014-08-16 20:52:06: dd failed'
2014-08-16 20:52:06: dd failed

Console log on the client:

LustreError: 3715:0:(vvp_io.c:1081:vvp_io_commit_write()) Write page 1213956 of inode ffff88007ccedb78 failed -28

Maloo report: https://testing.hpdd.intel.com/test_sets/acafc288-26a6-11e4-84f2-5254006e85c2

The failure was reported in ~~LU-3326~~ before. However, Lustre b2_5 build #80 already contained patch http://review.whamcloud.com/11425. So, more fixup is needed to resolve the failure.

Whamcloud Community JIRA

[LU-5526] recovery-mds-scale test failover_mds: dd: No space left on device