Bug 1034225

Summary:	[RFE] Abort if port 8702 (ajp proxy) is taken by a different process
Product:	Red Hat Enterprise Virtualization Manager	Reporter:	David Jaša <djasa>
Component:	ovirt-engine	Assignee:	Nobody <nobody>
Status:	CLOSED WONTFIX	QA Contact:
Severity:	low	Docs Contact:
Priority:	unspecified
Version:	3.3.0	CC:	acathrow, alonbl, didi, iheim, juan.hernandez, lpeer, oschreib, Rhev-m-bugs, sbonazzo, yeylon
Target Milestone:	---	Keywords:	FutureFeature, Improvement
Target Release:	---
Hardware:	Unspecified
OS:	Unspecified
Whiteboard:	integration
Fixed In Version:		Doc Type:	Enhancement
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2014-08-03 06:35:29 UTC	Type:	Bug
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:

Description David Jaša 2013-11-25 13:00:53 UTC

Description of problem:
During hunting some low-level bug, I managed to confuse init that the service is stopped or in "process dead but pid file exists" while some process escaped and it was still listening on port 8702. The result was that while engine appeared to work from cli (service ovirt-engine running etc.), the web ui threw nonsensical errors.
If engine aborts right after finding out about failure to bind to port 8702 (or any other port where the engine service listens), the error is trivial to pinpoint.

Version-Release number of selected component (if applicable):
3.3 (is24.1) / rhevm-backend-3.3.0-0.36.beta1.el6ev.noarch

How reproducible:
always

Steps to Reproduce:
1. make some process listen on port 8702 when the engine is stopped
2. start engine
3.

Actual results:
engine gets started

Expected results:
engine should fail to start

Additional info:

Comment 1 Sandro Bonazzola 2013-12-24 08:16:24 UTC

I'm not sure that port validation is something that should be done by service scripts...
It should be the engine process that should die if it can't bind the port. service scripts should already be able to detect it has not been started.
Or am I missing something?

Comment 2 Alon Bar-Lev 2013-12-24 08:32:56 UTC

jboss is started even if there are errors... these kind of errors are usually visible at server.log, unsure that this state is valid to invest extra work or code outside of jboss scope.

I am not sure how to instruct jboss to exit on such errors nor if it is simple.

Comment 3 Itamar Heim 2014-08-03 06:35:29 UTC

Closing old bugs. If this issue is still relevant/important in current version, please re-open the bug.