Bug 63645

Summary:	Cluadmin doesn't handle SIGTERM
Product:	Red Hat Enterprise Linux 2.1	Reporter:	Lon Hohberger <lhh>
Component:	clumanager	Assignee:	Jason Baron <jbaron>
Status:	CLOSED ERRATA	QA Contact:
Severity:	medium	Docs Contact:
Priority:	low
Version:	2.1	CC:	knoel
Target Milestone:	---	Keywords:	FutureFeature
Target Release:	---
Hardware:	i386
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	Enhancement
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2002-06-21 15:36:32 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:

Description Lon Hohberger 2002-04-16 19:30:09 UTC

Description of Problem:
Cluadmin does not exit cleanly when sent a TERM signal - it does not release the
cluster session lock.

Version-Release number of selected component (if applicable): 1.0.11-1

How Reproducible: Always.

Steps to Reproduce:
1. Run cluadmin.
2. Log in on another pty and run "killall -TERM cluadmin"
3. Run cluadmin again.

Actual Results:
Nasty message asking if you want to seize the cluster session lock when cluadmin
is run after it's been TERMinated.

Expected Results:
Clean reentry.

Additional Information:

Comment 1 Lon Hohberger 2002-06-21 15:36:26 UTC

Enhancements in pool.  As a side effect, I also changed the lack of handling of
Ctrl-C in the same manner, making the UI part of cluadmin *slightly* more sane -
any complete 'commit' operation succeeds, but I also note that signals are
ignored _during_ all operations involving writing the cluster database:

- Lock
- UnLock
- Write
- WriteFile

So, in a few instances, "Ctrl-C" will be ignored to prevent data corruption.