[LU-5622] copytool_cleanup function should check/wait for copytool death

• Type: Bug
• Resolution: Fixed
• Priority: Minor
• Fix Version/s: Lustre 2.7.0, Lustre 2.5.4
• Affects Version/s: Lustre 2.7.0, Lustre 2.5.3
• Labels: 22pl hsm
• Severity: 3
• Assignee: Bruno Faccini (Inactive)
• Reporter: Bruno Faccini (Inactive)

Description

copytool_cleanup() function of sanity-hsm test suite does not presently check/wait for effective copytool main+fork'ed processes death.

This can be a problem, since a lot of sanity-hsm sub-tests end with a copytool_cleanup() call when next sub-test starts with a copytool_setup() call, and may lead to situations/failures, like in the following auto-tests sessions :

https://testing.hpdd.intel.com/test_sets/2fb4cdd8-3aaa-11e4-a021-5254006e85c2
https://testing.hpdd.intel.com/test_sets/7ff7ee42-2fe2-11e4-9f89-5254006e85c2
https://testing.hpdd.intel.com/test_sets/1bc3d036-311c-11e4-b503-5254006e85c2
https://testing.hpdd.intel.com/test_sets/02b03f6c-e761-11e3-9c5d-52540035b04c

where next sub-test copytool_setup() attempt will not succeed to restart a new/fresh copytool (but only signal the current/old one from previous sub-test), finally leading to a test failure.

Issue Links

• is related to: LU-6203 sanity-hsm test 251: FAIL: Copytool failed to stop in 20s
• is related to: LU-7312 sanity-hsm: $? verification is not valid with set -e