Bug 1023638

Summary:	[RFE] Change location of volume configuration files
Product:	[Community] GlusterFS	Reporter:	Dean Bruhn <deanbruhn>
Component:	core	Assignee:	Vijay Bellur <vbellur>
Status:	CLOSED WONTFIX	QA Contact:
Severity:	unspecified	Docs Contact:
Priority:	unspecified
Version:	mainline	CC:	atumball, bugs, cww
Target Milestone:	---	Keywords:	RFE, Triaged
Target Release:	---
Hardware:	Unspecified
OS:	Unspecified
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2018-10-24 09:59:01 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:
Embargoed:

Description Dean Bruhn 2013-10-26 12:02:53 UTC

Description of problem:
Currently the gluster config files exist @ /var/lib/glusterd, the /var directory is prone to overruns, it might be a good idea to change the configuration files storage to /etc/glusterd to avoid this. 

Alternately the installation documentation should at least warn of this possibility and suggest keeping the configuration files on an alternate partition/storage location so the if the gluster log's become chatty, it doesn't compound an existing issue.  

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


How reproducible:


Steps to Reproduce:
1.
2.
3.

Actual results:


Expected results:


Additional info:

Comment 1 Harshavardhana 2013-10-29 16:54:59 UTC

(In reply to Dean Bruhn from comment #0)
> Description of problem:
> Currently the gluster config files exist @ /var/lib/glusterd, the /var
> directory is prone to overruns, it might be a good idea to change the
> configuration files storage to /etc/glusterd to avoid this. 
> 
> Alternately the installation documentation should at least warn of this
> possibility and suggest keeping the configuration files on an alternate
> partition/storage location so the if the gluster log's become chatty, it
> doesn't compound an existing issue.  

would you explain more about "overruns" ?

Comment 2 Dean Bruhn 2013-10-29 17:31:55 UTC

By overrun, I mean the log files can fill the partition up, and if the partition becomes full the configuration files become corrupted. I have had instances where a mnt having self healing issues has created a 20gb log file over night, and a rebalance log file has been 11gb or better. 

Both edge cases, but I have had a production system go down 3+ times because of the issue.

Comment 3 Joe Julian 2013-10-29 17:32:26 UTC

/var/log is a common management issue. You need to be aware of your logs and rotate and them as needed.

/var/lib is the appropriate directory for "application state" according to the Filesystem Hierarchy Standard and state data is what's in /var/lib/glusterd, not editable configuration files.

What I always do is make /var/log it's own partition to mitigate the issues you're concerned with.

Comment 4 Kaleb KEITHLEY 2015-12-08 12:21:38 UTC

Will we ever change this? If not it should be closed.

Comment 5 Amar Tumballi 2018-10-24 09:59:01 UTC

We will continue to use /var/lib/glusterd for the life of glusterd (v1). With glusterd2 project, the volume file store is changed to etcd service, which takes care of all this.