Modify update script to cover package configs updating
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Fuel for OpenStack |
Fix Released
|
High
|
Ivan Berezovskiy | ||
Mitaka |
Fix Released
|
High
|
Ivan Berezovskiy |
Bug Description
Some time ago the updating script was introduced [1]. It replaced old way when `fuel-bootstrap` was called.
We (fuel-qa team) had switched all our updating procedures to this new script. Our customers will use this script also.
When we updated the master node with this script we faced a problem with packets fuel-agent and fuel-bootstrap-cli and it's updated config which didn't update an old config on the master.
At the fuel-agent package spec we found `%config(noreplace) %{_sysconfdir}
The result of the update script is:
+ echo 'Master node update is <successfully complete> (log: /var/log/
Master node update is <successfully complete> (log: /var/log/
+ exit 0
But in fact we have an updated package fuel-bootstrap-cli and an outdated config `fuel_bootstrap
And additional work with rpm -U was needed to replace old config:
rpm -e fuel-bootstrap-cli
rpm -Uhv --force fuel-bootstrap-
So I suggest to add some checks and warning to the output to avoid misunderstanding of update status.
Or add some interactive questions asked what user should do with old configs if new ones are provided.
[1] https:/
tags: | added: swarm-blocker |
Changed in fuel: | |
status: | New → Confirmed |
tags: | added: area-library |
Changed in fuel: | |
assignee: | Fuel Sustaining (fuel-sustaining-team) → Ivan Berezovskiy (iberezovskiy) |
Changed in fuel: | |
assignee: | Ivan Berezovskiy (iberezovskiy) → Maksim Malchuk (mmalchuk) |
Changed in fuel: | |
assignee: | Maksim Malchuk (mmalchuk) → Ivan Berezovskiy (iberezovskiy) |
milestone: | 9.1 → 10.0 |
Changed in fuel: | |
status: | Fix Committed → Fix Released |
Fuel-sustaining team, is it safe to always update fuel_bootstrap_ cli.yaml config? If not, what is the solution you can suggest here?