Unsolved
This post is more than 5 years old
4 Posts
2
37309
How to get rid of the LUNZ devices?
We are an EMC shop... and any new disk allocations are either from CLARiiON or Symmetrix.
I see LUNZ devices when the AIX servers are zoned to the disk array for the first time. Once the real LUNs are allocated from the disk array we delete those LUNZ devices and run cfgmgr/powermt config to discover the LUNs. Some times the LUNZ's disappear and some times they stick around.
Why do these LUNZ devices stick around and How to get rid of these LUNZ devices? Has it got something to do with the way LUNs are allocated?
# lsdev -Cc disk | grep LUNZ
hdisk250 Available 11-T1-01 EMC CLARiiON FCP LUNZ Disk
hdisk251 Available 11-T1-01 EMC CLARiiON FCP LUNZ Disk
hdisk252 Available 12-T1-01 EMC CLARiiON FCP LUNZ Disk
hdisk253 Available 12-T1-01 EMC CLARiiON FCP LUNZ Disk
hdisk254 Available 21-T1-01 EMC CLARiiON FCP LUNZ Disk
hdisk255 Available 21-T1-01 EMC CLARiiON FCP LUNZ Disk
hdisk256 Available 22-T1-01 EMC CLARiiON FCP LUNZ Disk
hdisk257 Available 22-T1-01 EMC CLARiiON FCP LUNZ Disk
Please suggest or point me in the right direction.
Thanks
Maliboo74
89 Posts
1
January 3rd, 2013 09:00
Hello,
LUNZ devices will always be present when you make the first physical connection to the Clariion. Once you create and assign a lun with a host ID of zero then the LUNZ's can be removed.
dynamox
2 Intern
2 Intern
•
20.4K Posts
0
January 3rd, 2013 10:00
make sure your system admins do run cfgmgr or reboot the box until after you present the storage. LUNZ can be removed using rmdev -dl hdisk#
Maliboo74
89 Posts
1
January 3rd, 2013 10:00
Yes, it is best practice. The first lun created will default to zero. There are many articles about LUNZ behavior in the knowledgebase. Here is one example:
emc104353 "How do I change the LUN number of a CLARiiON device?"
In addition, take a look at the EMC Host Connectivity Guide for your particular OS for more details.
gshekar1
4 Posts
0
January 3rd, 2013 10:00
If we run cfgmgr after deleting the LUNZ; they are discovered again. So its apparent that the storage administrator did not allocate a LUN with Host ID=0.
A few more queries regarding this:
1. How does a UNIX admin determine if a LUN with Host ID=0 was allocated from the disk array?
2. Is it a documented / standard practice to allocate atleast one LUN with Host ID=0 from the disk array?
Thanks,
dynamox
2 Intern
2 Intern
•
20.4K Posts
0
January 3rd, 2013 10:00
In summary, this occurs because the arraycommpath setting of 1 creates a virtual LUN 0 for communication with the storage system.
A way of getting rid of LUNZ is to set arraycommpath to a value of 0. The arraycommpath option enables or disables a communication path from the server to the storage system.
This does not work on AX-series arrays. AX150 arrays are shipped with the failover/arraycommpath mode configured as 1 by default. It is not possible to change the failover mode using Navisphere Express or Navisphere Manager. Also, it is not possible to configure the failover mode and arraycommpath setting using CLI.
For CLARiiON AX-series you would need to use storagegroup command:
storagegroup sethost -arraycommpath
Caution! Do not issue the storagegroup command unless you know the appropriate value for the type of failover software running on the attached servers. The command is designed for transition from one type of failover software to another. Changing the arraycommpath mode to the wrong value will make the storage system inaccessible.
SKT2
2 Intern
2 Intern
•
1.3K Posts
0
January 8th, 2013 12:00
Do you need any other additional information on this now?
Jithu9288
61 Posts
0
December 7th, 2013 08:00
Hi Dynamox,
Though we have allocated lun to an AIX server in 0th position the server still sees the LUNZ device when it got rebooted. Arraycompath is enabled and failover mode is 3. is that okay to change the arraycompath to disabled so that the server wont see the lunz anymore or will that cause any issues?
root@******b01:/> lsdev -Cc disk |grep -i LUNZ
hdisk0 Defined 00-00-02 EMC CLARiiON FCP LUNZ Disk
hdisk2 Defined 00-00-02 EMC CLARiiON FCP LUNZ Disk
hdisk3 Defined 02-00-02 EMC CLARiiON FCP LUNZ Disk
hdisk115 Defined 02-00-02 EMC CLARiiON FCP LUNZ Disk
root@********b02:/> lsdev -Cc disk |grep -i LUNZ
hdisk2 Defined 01-00-02 EMC CLARiiON FCP LUNZ Disk
hdisk3 Defined 01-00-02 EMC CLARiiON FCP LUNZ Disk
hdisk114 Defined 03-00-02 EMC CLARiiON FCP LUNZ Disk
hdisk115 Defined 03-00-02 EMC CLARiiON FCP LUNZ Disk
hdisk178 Defined 03-00-02 EMC INVISTA FCP LUNZ Disk
hdisk181 Defined 03-00-02 EMC INVISTA FCP LUNZ Disk
Tandok
34 Posts
1
December 7th, 2013 09:00
Check if there are additional paths zoned, but registered.
Also get into Engineering mode(CtrlshiftF12, password: messner), go into the storage group>connect hosts>advanced. verify if all the paths have a tick against them.
LUNZ can be seen against any path connected, through which the luns are not visible.
Regards,
Karan
dynamox
2 Intern
2 Intern
•
20.4K Posts
1
December 7th, 2013 13:00
if you have multiple paths, using PowerPath or native MPIO ..you want to use failover mode 3 for proper NDU upgrades in the future. You need to figure out why you are seeing LUNZ. Check what Karan suggested, also try to remove LUNZ (rmdev) and reboot or cfgmgr.
anand_s_krishna
1 Message
0
December 9th, 2015 03:00
This KB should be useful in understand everything about LUNZ on IBM AIX hosts and how to remove them : https://support.emc.com/kb/7190