Bug 2055028

Summary:	Cluster pacemaker complains about a resource which is not there
Product:	Red Hat Enterprise Linux 9	Reporter:	lejeczek <peljasz>
Component:	pacemaker	Assignee:	Ken Gaillot <kgaillot>
Status:	CLOSED INSUFFICIENT_DATA	QA Contact:	cluster-qe <cluster-qe>
Severity:	high	Docs Contact:
Priority:	unspecified
Version:	CentOS Stream	CC:	bstinson, cluster-maint, jwboyer, tojeline
Target Milestone:	rc	Flags:	pm-rhel: mirror+
Target Release:	---
Hardware:	x86_64
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	If docs needed, set a value
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2022-04-19 16:27:47 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 lejeczek 2022-02-16 09:04:25 UTC

Description of problem:

Hi.
This is from: /var/log/pacemaker/pacemaker.log
...
Feb 16 09:01:30.977 swir.mine.private crm_resource        [3107823] (native_unpack) 	error: Resource GATEWAY:0 is of type systemd and therefore cannot be used as a promotable clone resource
Feb 16 09:01:30.977 swir.mine.private crm_resource        [3107823] (pe__create_clone_child) error: Failed unpacking resource GATEWAY
Feb 16 09:01:30.977 swir.mine.private crm_resource        [3107823] (unpack_resources) 	error: Ignoring <clone> resource 'GATEWAY-clone' because configuration is invalid
Feb 16 09:01:31.014 swir.mine.private crm_resource        [3107872] (clone_unpack) 	error: Ignoring promoted-max for GATEWAY-clone because anonymous clones support only one instance per node
Feb 16 09:01:31.014 swir.mine.private crm_resource        [3107872] (native_unpack) 	error: Resource GATEWAY:0 is of type systemd and therefore cannot be used as a promotable clone resource
Feb 16 09:01:31.014 swir.mine.private crm_resource        [3107872] (pe__create_clone_child) error: Failed unpacking resource GATEWAY

but, there is no such resource defined
-> $ pcs status --full | grep -i gateway
(blank)
-> $ pcs resource status | grep -i gateway
(blank)

thanks, L.


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

pacemaker-schemas-2.1.2-4.el9.noarch
pacemaker-libs-2.1.2-4.el9.x86_64
pacemaker-cluster-libs-2.1.2-4.el9.x86_64
pacemaker-cli-2.1.2-4.el9.x86_64
pacemaker-2.1.2-4.el9.x86_64

How reproducible:


Steps to Reproduce:
1.
2.
3.

Actual results:


Expected results:


Additional info:

Comment 1 Tomas Jelinek 2022-02-16 12:25:43 UTC

I think that this is due to a misconfiguration explained in the log you provided:
Resource GATEWAY:0 is of type systemd and therefore cannot be used as a promotable clone resource

Try running 'pcs resource config' and check if you can see the resource defined.

Comment 2 Ken Gaillot 2022-02-16 14:48:52 UTC

Hi,

Can you attach the CIB (or a pcs cluster report) from the time this happened? If you look in the logs on the cluster nodes, find the closest "saving inputs" message before the relevant time, and it will show a file name that is the CIB from that moment.

Comment 3 lejeczek 2022-02-16 15:49:34 UTC

Hi.
There is config for that resource, according to 'resource config' but the fact that 'cluster' does not show the resource in "usual" places made me decide it was a BZ.(no?)

regarding logs - perhaps a side "issue" - my /var/log/pacemaker/pacemaker.log grew freaking gigantic(>2GB with entries few months old), had to run logrotate manually and I wonder if logrotate's problem or ha-cluster's - I just thought I'd share here.

thanks, L.

Comment 4 Ken Gaillot 2022-02-17 16:00:28 UTC

(In reply to lejeczek from comment #3)
> Hi.
> There is config for that resource, according to 'resource config' but the
> fact that 'cluster' does not show the resource in "usual" places made me
> decide it was a BZ.(no?)

I would expect it to show up in pcs status. If you have the CIB from when it didn't, we can check whether it's a bug.

> 
> regarding logs - perhaps a side "issue" - my
> /var/log/pacemaker/pacemaker.log grew freaking gigantic(>2GB with entries
> few months old), had to run logrotate manually and I wonder if logrotate's
> problem or ha-cluster's - I just thought I'd share here.
> 
> thanks, L.

The logrotate configuration in /etc/logrotate.d/pacemaker should rotate if the log hits 100MB (or at least weekly). The cluster doesn't deal with log rotation directly so it would be a logrotate issue.

Comment 5 Ken Gaillot 2022-04-19 16:27:47 UTC

If this continues to be an issue, we can investigate further if we get a scheduler input with the issue