Unsolved
This post is more than 5 years old
5 Practitioner
•
274.2K Posts
0
5638
VPLEX VPD ID WWN question
IHAC with multiple VPLEX clusters…they want to use the Virtual-Volume WWN to determine which VPLEX the volume is come from. I could not find any reference to how to do this.
Example:
VPLEX FNM00144501082, Cluster-2, Geosync 5.4.1.1 connect to VMAX array
Claimed VDP ID is VPD83T360000970000192602932533031314438
Virtual-volume VDP ID is VPD83T3:60001440 000000103066e8dc75df 2b40
The host sees this as WWN: 60001440 000000103066e8dc75df 2b40
Breaking it down using naa-6 format
60 – naa format
001440 – Lets us know it is a VPLEX
00000010 – Vender definition
3066e8dc75df 2b40 – Vender specification (this is what I need to decipher)
Last octet (40) is the volume number
2nd to last octet (2b) is somehow associated with volume # but not specifically sure how
Do not know what the remaining octets are showing (3066e8dc75df), but they do change from VPLEX to VPLEX and I have noted in one occasion within the same VPLEX.
Can anyone help?
Ayrton_Senna1
1 Rookie
1 Rookie
•
62 Posts
0
December 23rd, 2015 13:00
Which hosts OS these VVOLS are presented to ? Have you had a chance to use INQ binary ? INQ has lot of options, and will give detailed info about vplex serial number with respect to VPDID
Anonymous
5 Practitioner
5 Practitioner
•
274.2K Posts
0
December 24th, 2015 06:00
I know we can identify them using INQ, however, it is not on every host, which is why the question comes back to the WWN and does it contain serial # information.
Chris Fielding
Cell: 540-735-4014
EMC²
Ayrton_Senna1
1 Rookie
1 Rookie
•
62 Posts
0
December 24th, 2015 10:00
I couldn't find a way to decipher this, alternative is to place inq utility in a common share(NFS or CIFS) and you can access INQ from each and every host. Less tedious and easy to implement in my opinion. I am assuming you guys need this during troubleshooting or renaming disks on host side. If I can know the exact purpose of this I can suggest a better solution.
Anonymous
5 Practitioner
5 Practitioner
•
274.2K Posts
0
December 28th, 2015 07:00
I appreciate you looking into this. I know there are methods available ,such as inq…however, the customer has a script they used for identifying and naming the volume in VMWare for datastore using the device WWN ID. Though this worked in the past for VMAX and VNX arrays, they are now trying to use the script for VPLEX and it does not work, since the naa-6 format for vendor section is not publicly defined.
Chris Fielding
Cell: 540-735-4014
EMC²
Ayrton_Senna1
1 Rookie
1 Rookie
•
62 Posts
0
December 28th, 2015 11:00
Thanks for explaining the situation. I am assuming 2 scenarios here.
1) Storage Admins who have access to VPLEX's want to decipher VPLEX serial num from VPDID
--Easiest way to do this is collect nightly dumps from all VPLEX's with list of all virtual vols. From this dump you can relate which Vvol belongs to which VPLEX serial num. This can be achieved with a kron job and unix shell script. I have dumps setup from my Vplex's and each one takes approximately 10 mins to collect. This dump can be integrated into your existing scripts to make your scripts work for both VMAX and VPLEX.
2) Second scenario would be the VMware admin is trying to figure out which naa id belongs to which VPLEX
-- For this I recommend integrating inq in a script on ESXi host or install VMware Vsphere Storage viewer which can be installed on Vcenter Server, this will give detailed info of how many storage arrays are connected and VPLEX serial number can be obtained from the properties tab of each datastore with this plugin.
Anonymous
5 Practitioner
5 Practitioner
•
274.2K Posts
0
December 29th, 2015 06:00
That was my thought and I actually did that…this works somewhat, but did find some particularities.
As an example the following two WWN IDs are from the same VPLEX:
60001440000000 10b 00ad5 7d145e 89 a9
60001440000000 10a 00ad5 b03d2e e0 92
I guess as time goes on I would eventually have enough imperical data to determine the differences, however, since EMC sets the vender information I was hoping that someone would know the format so there was a clear understanding of how the octets are derived…haven’t found that person yet
Chris Fielding
Cell: 540-735-4014
EMC²
1 Attachment
image001.png