-
Notifications
You must be signed in to change notification settings - Fork 262
Conversation
|
Can one of the admins verify this patch? |
1 similar comment
|
Can one of the admins verify this patch? |
tests/rspec/lib/container_linux.rb
Outdated
| require 'ssh' | ||
|
|
||
| SSH_CMD_CONTAINER_LINUX_VERSION = 'sudo cat /var/lib/update_engine/prefs/aleph-version' | ||
| SSH_CMD_CONTAINER_LINUX_VERSION = 'sudo cat /etc/os-release | grep VERSION_ID= | cut -d = -f 2' |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This change here works for other platforms?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
From my understanding Container Linux should be shipping with this identification file, which is expected by systemd so should be fine in any platform. This makes the test deterministic as otherwise the test might be run before update-engine generates the aleph-version file
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This is not correct. This change would revert a fix that was merged in #2093. /etc/os-release contains the current version when we are actually interested in the originally installed version of CL. This change would not make the test deterministic: what if this SSH command was executed on the node after it rebooted? This is actually what was consistently happening and causing the tests to fail. In any case, a possible change would be to:
if [ -f /var/lib/update_engine/prefs/aleph-version ] ; then
sudo cat /var/lib/update_engine/prefs/aleph-version
else
source /usr/share/coreos/release && echo "$COREOS_RELEASE_VERSION"
fiThere was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
"/etc/os-release contains the current version when we are actually interested in the originally installed version of CL" thanks for spotting that @squat
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
will rebase with #2193
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
2193 is merged please rebase and lets get this merged as well
|
adding smoke tests to run in azure and aws because saw some changes that might affect other platforms |
|
ok to test |
|
@enxebre got this error: |
|
@cpanato thanks for having a look, variables issue should be fixed now. |
555e75f to
103e576
Compare
Makefile
Outdated
| -e GOOGLE_CLOUD_KEYFILE_JSON \ | ||
| -e GCLOUD_KEYFILE_JSON \ | ||
| -e GOOGLE_PROJECT \ | ||
| -e tectonic_gcp_ssh_key \ |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This needs to be TF_VAR_tectonic_gcp_ssh_key
modules/gcp/network/loadbalancer.tf
Outdated
|
|
||
| resource "google_compute_forwarding_rule" "tectonic-api-external-ssh-fwd-rule" { | ||
| load_balancing_scheme = "EXTERNAL" | ||
| name = "tectonic-api-external-ssh-fwd-rule" |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
please remove tectonic- from the name for consistency with other platforms, e.g. https://github.com/coreos/tectonic-installer/blob/master/modules/aws/master-asg/master.tf#L24
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Will rebase an change this after #2167 gets in
modules/gcp/network/loadbalancer.tf
Outdated
| @@ -1,5 +1,6 @@ | |||
| resource "google_compute_target_pool" "tectonic-master-targetpool" { | |||
| name = "tectonic-master-targetpool" | |||
| name = "tectonic-master-targetpool" | |||
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
same as below
tests/rspec/lib/container_linux.rb
Outdated
| require 'ssh' | ||
|
|
||
| SSH_CMD_CONTAINER_LINUX_VERSION = 'sudo cat /var/lib/update_engine/prefs/aleph-version' | ||
| SSH_CMD_CONTAINER_LINUX_VERSION = 'sudo cat /etc/os-release | grep VERSION_ID= | cut -d = -f 2' |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This is not correct. This change would revert a fix that was merged in #2093. /etc/os-release contains the current version when we are actually interested in the originally installed version of CL. This change would not make the test deterministic: what if this SSH command was executed on the node after it rebooted? This is actually what was consistently happening and causing the tests to fail. In any case, a possible change would be to:
if [ -f /var/lib/update_engine/prefs/aleph-version ] ; then
sudo cat /var/lib/update_engine/prefs/aleph-version
else
source /usr/share/coreos/release && echo "$COREOS_RELEASE_VERSION"
fi|
Rebased, should be good now |
|
ok to test |
|
@cpanato, I saw this on the AWS VPC test: |
|
@enxebre looking good but we need to rebase now that the resource rename pr merged. |
8ea2670 to
c1214a7
Compare
|
@squat thanks, just rebased it |
|
With the latest rebase I'm seeing this happening occasionally when running the tests on gcp: haven't dug into it yet |
|
Yes does seem unrelated. I’d like to see if recent changes to master fix these flakes |
|
ok to test |
|
the issues in Azure is because the latest CL image is not there :/ |
*.tectonic.gcp.dev.coreos.systemneeds to be set to point to gcp NS