Bug 717858

Summary: [scalability] connectStorageServer: discovery command is sent on each login although portal is the same for all nodes
Product: [Retired] oVirt Reporter: Haim <hateya>
Component: vdsmAssignee: Dan Kenigsberg <danken>
Status: CLOSED WONTFIX QA Contact:
Severity: high Docs Contact:
Priority: unspecified    
Version: unspecifiedCC: abaron, amureini, bazulay, danken, dnaori, hateya, iheim, mgoldboi, yeylon, ykaul
Target Milestone: ---   
Target Release: 3.3.4   
Hardware: x86_64   
OS: Linux   
Whiteboard: storage
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2013-02-03 12:24:50 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: Storage RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Bug Depends On: 612978    
Bug Blocks:    
Attachments:
Description Flags
vdsm log none

Description Haim 2011-06-30 08:23:12 UTC
Created attachment 510609 [details]
vdsm log

Description of problem:

scenario: 

on connectStorageServer command, iSCSI, VDSM performs discovery command (iscsiadm -m discovery -t st -p <ip>:<port>) for each node, with regardless if portal is the same for all nodes.
this has scalability & performance affect. 

actual affect: connectStorageServer takes 4 minutes with 182 targets (in my case, 2 portals, each portal expose 91 targets:

[root@nott-vds2 ~]# grep "/sbin/iscsiadm -m discoverydb -t sendtargets -p 10.35.64.103:3260 --discover" /tmp/checkSesssion  |wc -l
91

[root@nott-vds2 ~]# grep "/sbin/iscsiadm -m discoverydb -t sendtargets -p 10.35.64.106:3260 --discover" /tmp/checkSesssion  |wc -l 
91


example: 

portal: 10.35.64.103

node: iqn.2011-06.com.redhat:test1
node: iqn.2011-06.com.redhat:test2

login: 

iscsadm -m discovery -t st -p 10.35.64.103:3260

iscsadm -m node -T iqn.2011-06.com.redhat:test1 -l 

iscsadm -m discovery -t st -p 10.35.64.103:3260 

iscsadm -m node -T iqn.2011-06.com.redhat:test2 -l 

VDSM should take into account portal address, and in case all nodes resides within 1 portal, skip discovery command

Comment 2 Itamar Heim 2013-02-03 12:24:50 UTC
Closing old bugs. If this issue is still relevant/important in current version, please re-open the bug.