[WFCORE-6815][Community] Utility to reload a Host Controller to a different stab… #570
Conversation
|
@yersan Yes, but as an FYI I'm on PTO for the a week starting Wednesday so if this is time sensitive, I might not be that helpful :) |
|
Hi @kabir / @jamezp Any chance to look at this proposal? The lack of ability to segregate the tests is blocking work for the conversion branches, since we need to properly avoid the execution of tests that are not meant for the default stability level. If possible, I would like to unlock this. Thanks! |
| - The operation will be registered at stability level 'community' at present. | ||
| - The operation is based on the reload operation, and adds a parameter called `stability`. This parameter takes the stability level we want to reload the server to. | ||
| - The operation also supports all the parameters supported by the standard reload operation, included the restart-servers parameter. | ||
| - Restarting a host controller with a different stability level without restarting the servers to get the value of the new stability level is not an useful operation, for simplicity and consistency with the reload operation, the restart-servers parameter will be allowed. |
There was a problem hiding this comment.
Should this say the "restart-servers parameter will not be allowed"? I'm asking because the first part of the sentence makes it not sound useful.
If the wording is intended, is the host controller meant to act as the minimum stability level for servers?
There was a problem hiding this comment.
Should this say the "restart-servers parameter will not be allowed"?
It is currently allowed, but it does not make sense to reload the host to a certain stability level and not restart your servers. If you do that, you could have your host running at one stability level, but your server running at another. I think that could be a problem. Maybe the simplest approach here would be to remove the restart-servers attribute from the reload-enhanced operation and restart the servers always when this operation is used.
I initially wanted to avoid these differences and assume you should know what you are doing when using the reload-enhanced, after all it is for our testsuite, That would bring more flexibility, but maybe that's not a good idea. WDYT?
is the host controller meant to act as the minimum stability level for servers?
The Host Controller sends its stability level to the managed servers with the managed server boot command, that's all. It dictates what stability level the managed server will run once it gets restarted. So, if you do not restart your servers, you could have the Host Controller running on preview and the managed servers still on community, I guess that could be problematic somehow, but I'm not sure.
Maybe we should start as restrictive as we can, e.g. remove the restart servers from reload enhanced and always restart them to avoid problems, and if there is a need, open it to allow restarting the server to get more flexibility.
There was a problem hiding this comment.
@yersan IMO then I think the restart-servers attribute should be removed from the new operation. I can understand not wanting the difference, However, attributes on operations that don't exist are ignored so it won't really hurt anything and makes it safer I think.
There was a problem hiding this comment.
@jamezp I've removed the restart-servers attribute from the reload-enhanced operation. Let me know how it looks like now
yersan
force-pushed
the
WFCORE-6815
branch
2 times, most recently
from
041b5ea to
a3c59e4
Compare
|
@kabir Let me know if there is anything else to add here, otherwise, approve if you agree, thanks! |
|
@kabir Friendly remainder, this one is waiting for your approval to get it merged, please take a look, thanks! |
…ility level in the testsuite Jira issue: https://issues.redhat.com/browse/WFCORE-6815
bstansberry
changed the title
…ility level in the testsuite
Jira issue: https://issues.redhat.com/browse/WFCORE-6815