[LU-6186] EL7 client sanity-hsm test_70: Failed to start copytool monitor on

• Status: Resolved
• Project: Lustre
• Affects Version/s: Lustre 2.7.0, Lustre 2.8.0
• Fix Version/s: Lustre 2.8.0
• Type: Bug
• Priority: Minor
• Reporter: Maloo
• Assignee: Bruno Faccini (Inactive)
• Resolution: Fixed

Environment

server: lustre-master build # 2830 RHEL6
client: EL7

• Severity: 3

Description

This issue was created by maloo for sarah <[e-mail]>

This issue relates to the following test suite run: https://testing.hpdd.intel.com/test_sets/659ff066-a496-11e4-a1ef-5254006e85c2.

The sub-test test_70 failed with the following error:

test failed to respond and timed out

test log

== sanity-hsm test 70: Copytool logs JSON register/unregister events to FIFO ========================= 00:43:52 (1422089032)
CMD: onyx-43vm5 pkill -INT -x lhsmtool_posix
CMD: onyx-43vm5 mktemp --tmpdir=/tmp -d sanity-hsm.test_70.XXXX
CMD: onyx-43vm5 mkfifo -m 0644 /tmp/sanity-hsm.test_70.74FK/fifo
CMD: onyx-43vm5 cat /tmp/sanity-hsm.test_70.74FK/fifo > /tmp/sanity-hsm.test_70.74FK/events & echo \$! > /tmp/sanity-hsm.test_70.74FK/monitor_pid
ps: write error: Bad file descriptor
 sanity-hsm test_70: @@@@@@ FAIL: Failed to start copytool monitor on onyx-43vm5 

Comments

[Andreas Dilger]

It looks like this might be a bug in sanity-hsm.sh:

copytool_monitor_setup() {
        :
        :
                ps -p $HSMTOOL_MONITOR_PDSH >&- ||
                        error "Failed to start copytool monitor on $agent"

According to bash(1) the use of >&- means:

       Each redirection that may be preceded by a file descriptor number may
       instead be preceded by a word of the form {varname}.  In this case, for
       each redirection operator except >&- and <&-, the shell will allocate a
       file descriptor greater than or equal to 10 and assign it to varname.
       If  >&- or <&- is preceded by {varname}, the value of varname defines
       the file descriptor to close.

Since the ps p option is using $HSMTOOL_MONITOR_PDSH there is no {varname} for the >& option. I don't know what this test is doing, but I suspect that the RHEL7 bash is being more strict about handling this error due to recent security bugs in bash.

[Bruno Faccini (Inactive)]

Seems that RHEL7 bash may require the filedescriptor number/variable to be specified now.
So

"ps -p $HSMTOOL_MONITOR_PDSH 1>&-"

should make it.

[Gerrit Updater]

Faccini Bruno ([e-mail]) uploaded a new patch: http://review.whamcloud.com/13857
Subject: LU-6186 tests: specify fd in bash close syntax
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: 4bec1be8287d5f68b6a85253a29ae66b0e170b90

[Gerrit Updater]

Bob Glossman ([e-mail]) uploaded a new patch: http://review.whamcloud.com/14084
Subject: LU-6186 tests: force test of LU-6186 on el7
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: a083876f739bc20eaafa7960c413cd6ec3cdaace

[Sarah Liu]

another instance: https://testing.hpdd.intel.com/test_sets/9dd2d66c-1311-11e5-8d21-5254006e85c2

[Gerrit Updater]

Oleg Drokin ([e-mail]) merged in patch http://review.whamcloud.com/13857/
Subject: LU-6186 tests: avoid errors in using >&- bash close syntax
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: 68e6eb752f0bbc6e400403a8be4ed671938af075

[Peter Jones]

Landed for 2.8