[LU-6095] racer.sh should propagate $TRUNCATE to remote clients

• Status: Resolved
• Project: Lustre
• Affects Version/s: Lustre 2.7.0
• Fix Version/s: Lustre 2.7.0
• Type: Bug
• Priority: Minor
• Reporter: John Hammond
• Assignee: John Hammond
• Resolution: Fixed
• Labels: racer
• Severity: 3

Description

Running racer with 2 clients I noticed that $TRUNCATE (from init_test_env()) was not defined when file_truncate.sh was executed. This needs to be added to the list of propagated variables in test_1():

        for rdir in $RDIRS; do
                do_nodes $clients "DURATION=$DURATION MDSCOUNT=$MDSCOUNT \
                                   $racer $rdir $NUM_RACER_THREADS" &
                pid=$!
                rpids="$rpids $pid"
        done

In file_truncate.sh we should test that $TRUNCATE is set, exists and is executable before continuing

DIR=$1
MAX=$2

while true; do
        file=$DIR/$((RANDOM % MAX))
        $TRUNCATE $file $RANDOM 2> /dev/null
done

The net result of this is a lot of copies of sleep, sleeping for $RANDOM seconds

root     13252     1  0 25226   548   1 10:59 ?        00:00:00   /mnt/lustre/racer/11 19675
root     13505     1  0 25226   544   1 10:59 ?        00:00:00   /mnt/lustre2/racer/11 21574
root     13815     1  0 25226   544   1 10:59 ?        00:00:00   /mnt/lustre2/racer/11 21401
root     14293     1  0 25226   548   1 10:59 ?        00:00:00   /mnt/lustre2/racer/10 10467
root     19131     1  0 25226   548   1 11:00 ?        00:00:00   /mnt/lustre/racer/13 12387
root     21592     1  0 25226   532   0 11:00 ?        00:00:00   /mnt/lustre2/racer3/16 4617
root     21860     1  0 25226   532   0 11:00 ?        00:00:00   /mnt/lustre/racer3/16 21599
root     27685     1  0 25226   544   0 11:00 ?        00:00:00   /mnt/lustre2/racer1/11 16802
root     31564     1  0 25226   544   1 11:00 ?        00:00:00   /mnt/lustre/racer1/2 11079
root      5097     1  0 25226   536   1 11:00 ?        00:00:00   /mnt/lustre2/racer2/13 25438
root     16030     1  0 25226   544   0 11:01 ?        00:00:00   /mnt/lustre2/racer2/14 21194
root     21334     1  0 25226   532   0 11:01 ?        00:00:00   /mnt/lustre/racer2/3 29716
root     25177     1  0 25226   536   1 11:01 ?        00:00:00   /mnt/lustre2/racer1/5 15303
root     11551     1  0 25226   532   1 11:02 ?        00:00:00   /mnt/lustre2/racer3/11 27347
root     12460     1  0 25226   544   1 11:02 ?        00:00:00   /mnt/lustre/racer2/2 26261
root     14345     1  0 25226   536   0 11:02 ?        00:00:00   /mnt/lustre2/racer3/7 16972
root     15573     1  0 25226   548   1 11:02 ?        00:00:00   /mnt/lustre2/racer2/15 13108
root     16148     1  0 25226   536   0 11:02 ?        00:00:00   /mnt/lustre/racer3/7 13826
root     17143     1  0 25226   544   1 11:03 ?        00:00:00   /mnt/lustre/racer1/3 15848
root     11075     1  0 25226   548   0 11:04 ?        00:00:00   /mnt/lustre/racer3/18 19787
root     27160     1  0 25226   548   1 11:07 ?        00:00:00   /mnt/lustre/racer2/5 21942
root     10983     1  0 25226   532   0 11:09 ?        00:00:00   /mnt/lustre2/racer1/16 31928

In turn this makes it impossible for racer to complete successfully.

The other subscripts should be audited for similar issues.

Comments

[Andreas Dilger]

It might also make sense for the racer/file_truncate.sh script to have a default value for $TRUNCATE:

TRUNCATE=${TRUNCATE:-truncate}

so that this script works in the absence of external environment variables?

[Gerrit Updater]

Jinshan Xiong ([e-mail]) uploaded a new patch: http://review.whamcloud.com/13501
Subject: LU-6095 tests: define TRUNCATE program for racer
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: 50d2e77bc8b9c3a35d069186c0f27d7ccf3a7fde

[Gerrit Updater]

Oleg Drokin ([e-mail]) merged in patch http://review.whamcloud.com/13501/
Subject: LU-6095 tests: define TRUNCATE program for racer
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: 26cf2a21efc668081a08b4efe0a8a719cb2df40c

[Peter Jones]

Landed for 2.7