ci.jpl: Increase wait job timeout to 4.5hrs
Taking a case of https://ci.trustedfirmware.org/job/tf-m-nightly/1194/consoleFull,
14 of 638 jobs were cancelled due to older timeout expiration (13000s = ~3.6hrs).
There's no sign there that the jobs were stuck or something. There're just too
many jobs to run, and we need to give them reasonable timeframe to finish. If we
don't, we actually don't run some share of testsuite, and may miss issues creeping
in the codebase. At 4.5hrs = 16200s, we give 16200/638 ~ 25s for a job to finish,
so again, no indication that we have any hanging or slowdown in the job processing
(a random job of course takes much longer than 25s to finish, that figure takes into
account parallelization of running among several test devices).
Signed-off-by: Paul Sokolovsky <paul.sokolovsky@linaro.org>
Change-Id: I7be24735a199337bb0ac35927b2416963fa0848b
diff --git a/jenkins/ci.jpl b/jenkins/ci.jpl
index 7f6f2df..d0ee119 100644
--- a/jenkins/ci.jpl
+++ b/jenkins/ci.jpl
@@ -368,7 +368,7 @@
withCredentials([usernamePassword(credentialsId: env.LAVA_CREDENTIALS, passwordVariable: 'LAVA_TOKEN', usernameVariable: 'LAVA_USER')]) {
output = sh(script: """./tf-m-ci-scripts/lava_helper/lava_wait_jobs.py --job-ids ${all_jobs.join(",")} \
--lava-url ${env.LAVA_URL} --lava-user ${LAVA_USER} --lava-token ${LAVA_TOKEN} \
- --artifacts-path cfgs --lava-timeout 13000 \
+ --artifacts-path cfgs --lava-timeout 16200 \
""", returnStdout: true).trim()
println("--- output from lava_wait_jobs.py ---")
println(output)