It was discovered that healthchecks aren't really reliable because they
aren't strict enough.
The current patch adds the "standard" options in order to ensure we
actually catch errors soon enough in order to return the actual state of
the checked element.
It also requires a small change for the healthcheck_port() function,
since the "piping" returned a 141 code instead of 0 due SIGPIPE being
sent at some point[1].
It also depends on two other changes, in order to ensure we won't get
any "sudo" issues inside the checks (here again, healthcheck_port is
tricky).
Reviewed: https:/ /review. opendev. org/704651 /git.openstack. org/cgit/ openstack/ tripleo- common/ commit/ ?id=dddf584505b b24c53ba9d5fd92 396695dc7a540c
Committed: https:/
Submitter: Zuul
Branch: stable/train
commit dddf584505bb24c 53ba9d5fd923966 95dc7a540c
Author: Cédric Jeanneret <email address hidden>
Date: Wed Jan 22 16:23:24 2020 +0100
Make healthchecks more strict
It was discovered that healthchecks aren't really reliable because they
aren't strict enough.
The current patch adds the "standard" options in order to ensure we
actually catch errors soon enough in order to return the actual state of
the checked element.
It also requires a small change for the healthcheck_port() function,
since the "piping" returned a 141 code instead of 0 due SIGPIPE being
sent at some point[1].
It also depends on two other changes, in order to ensure we won't get
any "sudo" issues inside the checks (here again, healthcheck_port is
tricky).
[1] https:/ /stackoverflow. com/questions/ 19120263/ why-exit- code-141- with-grep- q www.tldp. org/LDP/ lpg/node20. html
http://
Depends-On: https:/ /review. opendev. org/703818 /review. opendev. org/703816 dc9c6ede46a939d 1853204f896 /bugzilla. redhat. com/show_ bug.cgi? id=1794044 6b775fd129e3a09 970862755b)
Depends-On: https:/
Change-Id: Iada9fb49881c8e
Closes-Bug: #1860556
Related: https:/
(cherry picked from commit ba02b0a23582b28