Unsolved
This post is more than 5 years old
99 Posts
0
1669
MDS scsi-target discovery
We've reinstalled some new Ciscso MDS switches in our fabric last week.
In our CX4's we had suddenly some initiators logging in which can't be found in the fcns database.
After some research it turned out that the MDS can do a scsi-target discovery using a combination of the switch vsan wwn and OS pwwn.
switch# show scsi-target pwwn
------------------------------
OS PWWN
------------------------------
WIN 24:91:00:05:30:00:2a:1e
AIX 24:92:00:05:30:00:2a:1e
SOL 24:93:00:05:30:00:2a:1e
LIN 24:94:00:05:30:00:2a:1e
HP 24:95:00:05:30:00:2a:1e
We've disbabled scsi-target discovery but we're still seeing the VSAN WWN:WIN PWWN logged in into our CX4.
We're not going to use this feature, is there any other way to keep the switch from logging into the CX4.
John
dynamox
2 Intern
2 Intern
•
20.4K Posts
0
October 28th, 2010 14:00
i wonder if iSCSI service running on the switch logs-in as initiator to the array, can you disable iSCSI and see if it logs out ? ( no iscsi enable)
ramkm
3 Posts
0
October 29th, 2010 01:00
restart the management server or SP A , SP B and check. Hope this could solve the issue.
jgrinwis
99 Posts
0
October 29th, 2010 04:00
Host are not registered, only logged in via the FC ports.
Not sure if disabling the iscsi part would solve the MDS from logging in via FC.
ramkm
3 Posts
0
October 29th, 2010 04:00
did try by deregister those initiators
dynamox
2 Intern
2 Intern
•
20.4K Posts
0
October 29th, 2010 04:00
that will not do anything, they will log back in.
ramkm
3 Posts
0
October 29th, 2010 05:00
Try rebooting of SAN switch. this wil referesh the FC login
dynamox
2 Intern
2 Intern
•
20.4K Posts
0
October 29th, 2010 05:00
reboot ? really ? do you realize what you are suggesting ? Reboot the switch fabric.
hersh1
197 Posts
0
October 29th, 2010 05:00
Take a look at this thread: https://community.emc.com/thread/94957?start=0&tstart=0
We've had this show up in our environment a couple different times and the way I've solved it is by registering the WWNs then switching to engineering mode and deregistering them. I don't think we have had them come back for a couple years now.
dynamox
2 Intern
2 Intern
•
20.4K Posts
0
October 29th, 2010 05:00
i have this issue with Symmetrix, would rather solve it on the switch side. I will try disabling iscsi service on one of the switches.
jgrinwis
99 Posts
0
October 29th, 2010 05:00
there should be another way to solve this, we're not going to reload our production MDS to just get rid off a initiator login.
jgrinwis
99 Posts
0
October 29th, 2010 06:00
If I check https://community.emc.com/thread/94957?start=0&tstart=0 looks like this issue might also be caused by fcping/FabricManager.
I've disabled the scsi discovery service on one of the switches, so when I register and deregister the initiator, one of the ghost wwn's should disappear, keep you informed.
dynamox
2 Intern
2 Intern
•
20.4K Posts
0
October 29th, 2010 07:00
disabling iSCSI service on the module did not work, i have 9222i that are iSCSI capable so was hoping that would cure it.
jgrinwis
99 Posts
0
November 1st, 2010 07:00
Registered the initiators, deregistered them again. After deregistering them, the intiators are no longer logged in.
We're running flare .29, did a "update now" a couple of times and initiators not seen anymore.
On one of the switches we did a "discovery scsi-target" for local and remote and it's seeing the LUNZ's. Did an "update now" and a "refresh all" but switch is not logging in anymore.
But when we issue a "fcping" we're seeing the switch logging in again over the port we're trying to fcping. So looks like fcping is causing this issue. Read that some people had this same issue when starting up fabric mananger, so fabric manager might issue an fcping during startup, causing these switch pwwn's to login to the different fc ports.