Bug 2059660

Summary:	[RFE]: Have Candlepin peform jobs in batches as opposed to all
Product:	Red Hat Satellite	Reporter:	Taft Sanders <tasander>
Component:	Candlepin	Assignee:	satellite6-bugs <satellite6-bugs>
Status:	CLOSED WONTFIX	QA Contact:	Vladimír Sedmík <vsedmik>
Severity:	high	Docs Contact:
Priority:	unspecified
Version:	6.9.6	CC:	bcourt, nmoumoul, redakkan, sriley
Target Milestone:	Unspecified	Keywords:	FutureFeature, Triaged
Target Release:	Unused
Hardware:	x86_64
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	If docs needed, set a value
Doc Text:		Story Points:	---
Clone Of:
Clones:	2069126 (view as bug list)		Environment:
Last Closed:	2023-10-06 10:59:23 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:
Bug Depends On:	2069126
Bug Blocks:

Description Taft Sanders 2022-03-01 16:30:32 UTC

Description of problem:
It has been identified that the following candlepin jobs will run with very long times and potentially fail with memory issues on tomcat:

ExpiredPoolsCleanupJob
RegenEnvEntitlementCertsJob

Version-Release number of selected component (if applicable):
Satellite 6.9.6-1
candlepin-3.1.28-1

How reproducible:
Every time

Steps to Reproduce:
1. Have a Satellite with 54,000 hosts with 1 subscription expire at the same time.
2. Watch tomcat usage reach 24GB+ memory usage
3.

Actual results:
Multiple attempts to complete the job as it has a restart if memory issues are encountered.
Long run times for these jobs with thousands of hosts

Expected results:
The job should be run in smaller batches assumed from a value in the configuration file.
IE:
while ExpiredPoolsExist:
  PoolBatchSize = config.parser(candlepin.poolbatchsize)
  if PoolBatchSize = '':
    set PoolBatchSize = 2000
  ExpiredPoolsCleanupJob(ExpiredPoolsList[:PoolBatchSize])

Additional info:
Configuring this for large deployments would also be needed to be added for the performance tuning guide as more memory is allocated to tomcat to process larger batches.

Comment 3 Nikos Moumoulidis 2022-03-21 11:20:44 UTC

Hi Taft,

Can you confirm whether Entitlement mode (versus SCA) was used in this deployment?

Thanks,
Nikos

Comment 4 Taft Sanders 2022-03-24 14:14:46 UTC

(In reply to Nikos Moumoulidis from comment #3)
> Hi Taft,
> 
> Can you confirm whether Entitlement mode (versus SCA) was used in this
> deployment?
> 
> Thanks,
> Nikos

I can confirm SCA was NOT enabled on the account/Satellite.

Comment 11 Nikos Moumoulidis 2023-10-06 10:59:23 UTC

While this is a valid issue, it only affects very large environents(in terms of numbers of hosts) in Entitlement Mode, and would require a high effort fix.
With Simple Content Access mode being the preferred method for accessing content, we have focused our efforts on that and decided not to fix this.