Bug 1321720

Summary:	getting unknown host when we ping CNAME after the application restart
Product:	Red Hat Enterprise Linux 6	Reporter:	Jayaraj <jdeenada>
Component:	glibc	Assignee:	Carlos O'Donell <codonell>
Status:	CLOSED WONTFIX	QA Contact:	qe-baseos-tools-bugs
Severity:	high	Docs Contact:
Priority:	urgent
Version:	6.6	CC:	ashankar, cww, fweimer, jdeenada, mnewsome, pfrankli
Target Milestone:	rc
Target Release:	---
Hardware:	x86_64
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2016-08-09 18:02:26 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:
Bug Depends On:
Bug Blocks:	1269194

Description Jayaraj 2016-03-29 03:02:32 UTC

Description of problem:

Issue: After restarting an application, when we ping it with CNAME, we are getting unknown host error for around 10-15 mins. nslookup is able to resolve during this time. After 10-15 mins, ping is also working. This happens everytime the application is restarted.

However if we bring down nscd, ping works fine without any error. 
if the customer use A record, then agents are able to communicate with OMS immediately after the application restart. Only when they use the CNAME we are facing this issue.

The package version installed:

nscd-2.12-1.149.el6_6.5.x86_64 

glibc-2.12-1.149.el6_6.5.i686
glibc-2.12-1.149.el6_6.5.x86_64 
glibc-common-2.12-1.149.el6_6.5.x86_64 
glibc-devel-2.12-1.149.el6_6.5.i686
glibc-devel-2.12-1.149.el6_6.5.x86_64
glibc-headers-2.12-1.149.el6_6.5.x86_64


Unable to reproduce the issue, however customer did a workaround the problem by lowering the TTL of CNAME record from 900sec to 20sec same as A record TTL. 
This proves that the issue was with NSCD, caching incorrectly similar to the issue documented in https://access.redhat.com/solutions/91233 .

Since the issue is occurring in RHEL 6.6 systems running  "glibc-2.12-1.149.el6.x86_64" (this version is already in a higher than the version that was first resolved as mentioned in the above said KB), i need some investigation. Could you please check if the bug mentioned before has re-surfaced or if this is a new bug?

Comment 11 Jayaraj 2016-04-05 02:02:07 UTC

Customer has asked time to simulate the issue and come back with tcpdump output as they currently applied a workaround to resolve the issue by bringing down the TTL value of CNAME to match with A record. I will keep the BZ updated as soon as I have the update from the customer.