Bug 429248

Summary: rgmanager restart previously stopped services
Product: Red Hat Enterprise Linux 5 Reporter: Fabio Massimo Di Nitto <fdinitto>
Component: rgmanagerAssignee: Lon Hohberger <lhh>
Status: CLOSED UPSTREAM QA Contact:
Severity: medium Docs Contact:
Priority: low    
Version: 5.3CC: cluster-maint
Target Milestone: rc   
Target Release: ---   
Hardware: All   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2008-01-18 19:36:51 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Attachments:
Description Flags
make help not suck none

Description Fabio Massimo Di Nitto 2008-01-18 09:09:44 UTC 
Description of problem:

rgmanager restart previously stopped services

Version-Release number of selected component (if applicable):

rgmanager from cvs HEAD. mostlikely applies to other released branches as well.

How reproducible:

always

Steps to Reproduce:
1. test cluster as described in the wiki
2. start rgmanager all nodes but one
3. stop one service (clusvcadm -s service_128)
4. start rgmanager on the only node that did not start previously

Actual results:

the service is started again.

Expected results:

the service should not start as it was previously stopped manually.
Comment 1 Lon Hohberger 2008-01-18 15:54:02 UTC 
Stop = "stop until next transition of (service | rgmanager) then re-evaluate"

Typically, rgmanager moves services to this state if a dependency is unsatisfied
or restricted-failover domain is offline.  Every service state change or node
state change triggers an evaluation to see if the service has become runnable.

Disable = "stop until administrator intervention"

Service is not restarted.  Unless re-enabled by an administrator with -e (or the
admin did an explicit enable in a RIND event script, which I consider bad form,
but is legal)


Try - clusvcadm -d instead of clusvcadm -s?
Comment 2 Fabio Massimo Di Nitto 2008-01-18 18:07:27 UTC 
Ok in this case it makes perfect sense and this is not a bug.

For some reason I would expect a manual stop to be an alias for disabled or at
least this is how i read it via clusvcadm -h

Fabio
Comment 3 Lon Hohberger 2008-01-18 19:35:57 UTC 
Created attachment 292188 [details]
make help not suck
Comment 4 Lon Hohberger 2008-01-18 19:36:51 UTC 
Moved to upstream cvs