[fuel-qa] devops fails to revert environment with one isolated node and external ntp server configured
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Fuel for OpenStack |
Fix Committed
|
Undecided
|
Vladimir Khlyunev |
Bug Description
Detailed bug description:
While devops revert the environment, he tries to sync time via ssh and ntpdate on each nodes in the environment.
The ntp serrver was choosed as first ntp server entry in /etc/ntp.conf
On the isolated node, we haven't access to public internet network, and ntpdate returns error code
Steps to reproduce:
1. Install fuel master 8.0 with ha openstack cluster
2. Install octane
3. Create backup
4. Reinstall fuel master node with 9.x release
5. Install octane
6. Restore
7. Start upgrade procedure via octane
8. ...
9. octane upgrade-node --isolated
10. Create snapshot
11. Revert snapshot
Expected results:
Reverting snapshot was successfully
Actual result:
Traceback (most recent call last):
File "/var/lib/
compatabili
File "/var/lib/
func()
File "/var/lib/
func(
File "/var/lib/
result = func(*args, **kwargs)
File "/var/lib/
self.
File "/var/lib/
self.
File "/var/lib/
new_time = sync_time(
File "/var/lib/
return func(*args, **kwargs)
File "/var/lib/
g_ntp.
File "/var/lib/
ntp.
File "/var/lib/
self.
File "/var/lib/
raise TimeoutError(
TimeoutError: Failed to set actual time on node u'slave-03'
Reproducibility:
100%
Workaround:
Before run upgrade-node --siolated, changed in fuel-qa test ntp server to the fuel master ip /etc/ntp.donf on target node
After upgrade, revert the changes in /etc/ntp.conf
OR
before running the command, make check, if node is in isolated mode, run ntpdate command with fuel master ip argument master
Impact:
Upgrade procedure
Description of the environment:
fuel-octane-
Additional information:
TimeoutError: Failed to set actual time on node u'slave-03'
Fix proposed to branch: stable/7.0 /review. openstack. org/366744
Review: https:/