Bug 2177191

Summary:	delays seen while sdap_sudo_refresh_done is running
Product:	Red Hat Enterprise Linux 8	Reporter:	mpanaous
Component:	sssd	Assignee:	Pavel Březina <pbrezina>
Status:	NEW ---	QA Contact:	sssd-qe
Severity:	medium	Docs Contact:
Priority:	unspecified
Version:	---	CC:	aboscatt, ipedrosa, pbrezina, rmarigny, thalman
Target Milestone:	rc	Keywords:	Triaged
Target Release:	---	Flags:	mpanaous: needinfo? (rmarigny)
Hardware:	Unspecified
OS:	Unspecified
Whiteboard:	sync-to-jira
Fixed In Version:		Doc Type:	If docs needed, set a value
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:		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:

Comment 6 Pavel Březina 2023-07-21 10:40:29 UTC

The sudo code is pretty much the same on RHEL-7 and RHEL8, since it is quite fast on RHEL-7 I wonder if there was any change in ldb (the library that we use for cache database) that would diminish the index performance. Tomas, I think you have some internal knowledge of ldb, can you please check?

Comment 8 Tomas Halman 2023-08-17 12:15:30 UTC

In RHEL7 we have version 1.5, in RHEL8 there is 2.6. The difference is significant but at first glance I do not see any obvious reasons why it is slower. 

In 2.6 is an option that allows to increase in-memory cache for indexes and that may help during update (it does not explain why it is so much slower now) but the commit message mentions huge
performance penalty if the cache is small and hence not used. Unfortunately there is no way to set this option in SSSD configuration currently.

I may run some experiments if we obtain the cache files, or we can create a scratch build with this memory cache option set to higher value just for the sake of investigation.