[LU-289] script error in oos.sh and oos2.sh

• Type: Bug
• Resolution: Fixed
• Priority: Trivial
• Affects Version/s: Lustre 1.8.8, Lustre 1.8.6
• Severity: 4
• Assignee: Jian Yu
• Reporter: Niu Yawei (Inactive)

Description

When I run sanityn.sh in my virtual machine, I found that the space is not freed after test_15 (which calls oos2.sh to fill up OST), that result in some of the following test failed for ENOSPC.

The reason of oos2.sh/oos.sh doesn't free space is that the output of 'dd if=/dev/zero of=$OOS count=$(($ORIGFREE + 100)) bs=1k 2> $LOG' is:

dd: writing `/mnt/lustre/oos': No space left on device
dd: closing output file `/mnt/lustre/oos': Input/output error

but not not the scprit exepected:

dd: writing `/mnt/lustre/oos': No space left on device
247330+0 records in
247329+0 records out
253264896 bytes (253 MB) copied, 121.715 seconds, 2.1 MB/s

so the following line of the script is broken, and the script exit without cleanup the test file.

RECORDSOUT=$((`grep "records out" $LOG | cut -d+ -f 1` + \
`grep "records out" $LOG2 | cut -d+ -f 1`))

I tried the dd command with 'oflag=direct', then the output will same as what we expected, but direct io will make the test slower, not sure if it's a good solution.

I think because the test_15 is always skipped (for too much free space on testing system), we never hit this issue in out daily test.

Issue Links

• Trackbacks: Lustre 1.8.6-wc1 release testing tracker Lustre 1.8.6wc1 RC1 Tag: v186RC1 Created Date: 20110610 RC1 was DOA due to a build failure related to tag name LU408
• Trackbacks: Lustre 1.8.8-wc1 release testing tracker Lustre 1.8.8wc1 RC1 Tag: v188WC1RC1 Build:
• Trackbacks: Lustre 1.8.x known issues tracker While testing against Lustre b18 branch, we would hit known bugs which were already reported in Lustre Bugzilla https://bugzilla.lustre.org/. In order to move away from relying on Bugzilla, we would create a JIRA