Bug 1839660

Created attachment 1691920 [details]
Trace Report showing query break down upon Edit Cluster via UI

No resource saturation occurs on Engine during this 'slow edit cluster' activity.
Open Edit Clusters, then press OK  (nothing is changed) total time: 117s
The cause of the long duration is /ovirt-engine/webadmin/GenericApiGWTService 

Of the total 117s, 57s are spent on 41,618 queries.


Breakdown:	total (ms)	count
http request   117,775.4	  1
jdbc query      57,447.5	41,618
jdbc commit      3,156.2	2,025
wait on future   1,719.7	187
jdbc get conn      655.9	40,033
logging            271.1	2,655

The following queries are executed more than 1000 times each

select * from getquotabyquotaguid(?)
select * from getdisksvmguid(?, ?, ?, ?)
select * from getvmbyvmguid(?, ?, ?)
select * from getvmdevicebyvmidtypeanddevice(?, ?, ?, ?, ?)
select * from getclusterbyclusterid(?, ?, ?)
select * from getvmnetworkinterfaceviewbyvmid(?, ?, ?)
SELECT NULL AS PROCEDURE_CAT, n.nspname AS PROCEDURE_SCHEM, p.proname A ... balloonenabled' ORDER BY PROCEDURE_SCHEM, PROCEDURE_NAME, p.oid::text
select * from getvdsbyvdsid(?, ?, ?)
select * from getvmdevicebyvmidandtype(?, ?)
select * from getvmdevicebyvmid(?, ?, ?)
SELECT n.nspname,p.proname,p.prorettype,p.proargtypes, t.typtype,t.typr ... LIKE 'ismemballoonenabled' ORDER BY n.nspname, p.proname, p.oid::text
select * from getvmtemplatebyvmtguid(?, ?, ?)
select * from getsnapshotbyvmidandtype(?, ?, ?, ?)
{call updatevmstatic(?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ...  ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)}
select * from getdiskvmelementsforvm(?, ?, ?)
select * from getstorage_poolbyid(?, ?, ?)
{call insert_entity_snapshot(?, ?, ?, ?, ?, ?, ?, ?, ?)}
{call insertauditlog(?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)}
select * from isvmiconexist(?)
{call incrementdbgeneration(?)}
{? = call ismemballoonenabled(?)}
select * from getquotaclusterbyquotaguid(?)
select * from getsnapshotsbyvmsnapshotid(?)
select * from getquotastoragebyquotaguid(?)
select * from getvminitbyvmid(?)
select * from getnumanodebyvmid(?)
select * from getvminterfacesbyvmid(?)
select * from getallassignednumanodeinfomation()


Please see attached: Cluster_Edit_GWT_trace.zip and open the resulting html file to see the full list of queries, and number of executions, and rows returned by query.

This issue maybe related to: https://bugzilla.redhat.com/show_bug.cgi?id=1811866  which has to do with many networks and Cluster UI view.

clicking OK does update all the VMs so naturally there will be 1000+ of each query
doesn't sound too interesting since it still works fine. 2 mins is not too bad

(In reply to Michal Skrivanek from comment #3)
> clicking OK does update all the VMs so naturally there will be 1000+ of each
> query

Is it required to update all VMs, even nothing is changed?

Tested again on Red-01  in order to reproduce.

version 4.4.2 :
       rhv-release-4.4.2-3-001.noarch
       vdsm-4.40.25-1.el8ev.x86_64
       L0_Group_0 with 1350 VMs 

didnt reproduce..

Yeah, triggering update-cluster with no change doesn't trigger update VMs/templates anymore

This bugzilla is included in oVirt 4.4.5 release, published on March 18th 2021.

Since the problem described in this bug report should be resolved in oVirt 4.4.5 release, it has been closed with a resolution of CURRENT RELEASE.

If the solution does not work for you, please open a new bug report.