Change wait in stop_hana crash #19228
Conversation
|
Great PR! Please pay attention to the following items before merging: Files matching
This is an automatically generated QA checklist based on modified files. |
Change the code to not only rely on wait_for_ssh but to polling `is-system-ready`.
mpagot
force-pushed
the
crash_stop_hana_wait_ssh
branch
from
a3dfe64
to
9e93106
Compare
| @@ -335,18 +335,28 @@ sub stop_hana { | |||
| if ($args{method} eq "crash") { | |||
| # Crash needs to be executed as root and wait for host reboot | |||
| $self->{my_instance}->wait_for_ssh(timeout => $timeout); | |||
There was a problem hiding this comment.
If I can add a kind suggestion, this call is not clear here if is waiting for some starting system; but in the case the ssh sut is wellknown to be already up in this point, the wait call could even be removed.
| sleep 60; | ||
| my $start_time = time(); | ||
| while ((time() - $start_time) < 60) { | ||
| $out = $self->{my_instance}->run_ssh_command( | ||
| cmd => "sudo systemctl is-system-running", | ||
| timeout => 5, | ||
| proceed_on_failure => 1); | ||
| last if ($out =~ m/running/); | ||
| sleep 10; | ||
| } | ||
|
|
There was a problem hiding this comment.
I think it not neeed here, you already have it in the next wait_for_ssh call body. May be an explicit ssh start is missing here after down state, instead ?
Try to check the sub softreboot() in instance.pm if could be useful to you.
There was a problem hiding this comment.
I mean, the full block 348 to 357 IMHO is a duplication, being already executed in L359 wait_for_ssh.
| sleep 10; | ||
| $out = $self->{my_instance}->wait_for_ssh(timeout => 900); | ||
|
|
||
| sleep 60; |
There was a problem hiding this comment.
As we are smartly waiting in the lines below, I'd drop this sleep 60 (and if needed, incorporate the time in the while sentence below as:
my $start_time = time();
whilte ((time() - $start_time) < 120) {
There was a problem hiding this comment.
"smart wait" in line below is not so smart. It keeps to have trouble with remote VM reboot. This sleep is to hope at least to enter in the while loop when the reboot is at least started.
| sleep 10; | ||
| } | ||
|
|
||
| $out = $self->{my_instance}->wait_for_ssh(timeout => 10); |
There was a problem hiding this comment.
I'm with @m-dati here: either we skip using publiccloud::instance::wait_for_ssh in L359 because it causes issues in our scenarios, or we implement something similar as in L348-L358. I don't see the need to keep both.
Change the code to not only rely on wait_for_ssh
Verification run:
sle-15-SP5-Azure-SAP-PAYG-Incidents-x86_64-Build:33599:rpm-SAPHanaSR-ScaleUp-PerfOpt-spn@az_Standard_E8s_v3
Error: waiting for Virtual Machine "Virtual Machine)sle-15-SP4-HanaSr-Azure-Byos-x86_64-Build15-SP4_2024-05-03T04:03:15Z-hanasr_azure_test_sapconf_msi@64bit
dieand that is not ok http://openqaworker15.qa.suse.cz/tests/282110#step/Crash_site_b-primary/425 🍅