Allocation queries

A good strategy moving forward might be to use the "Volume Identifier" field on your VMAX TDEVs to name them appropriately and rename the LUN itself on CLARiiON.  For instance, you could use "Server1_C" and "WebServer2_F" as useful names.  It might be time-consuming at first, but it could save you time down the road.

hey Karl , how do I set the Volume Identifier field 

Hi,

If you have access to the host in question, you can download and execute the inq utility to resolve the relationships between host devices and VMAX or Clariion/CX/VNX devices. You can get inq at ftp://ftp.emc.com/pub/symm3000/inquiry/v7.6.2.0. There are a bunch of different CLI arguments -- just run it with a -h to see the list of options.

Thanks,

- Sean

apparently my admins aint user friendly lol . I will defnitely try the inq utility .

cheers aran

the admins are repulsive when it comes to installing any s/w on their systems , they go crazy .

I can probably suggest the inq utility  when there is a new server build in our environment thats san connected  , better done during the initial set up  .

Thanks for the info Sean .

Cheers !!

I second the use of inq on the host, makes local device discovery quick and easy. I have found that the sysadmins/dbas like it as well and are willing to run it and send me the output once I show them how to use it.

one nice thing about inq is you don't have to install anything, just drop the executable in a directory and call it from the cli.

the volume identfier or 'device name' can be set as follows with the cli (or in Unisphere):

set dev devcie_name= ;

my personal experience with device names on a symm is that they aren't viewable in the output you may want to see them, so i don't bother with them anymore. dynamox may have some good examples about how he uses them.

Hi Aran,

never considered using that option because most of our boxes are either running SE, inq or we simply rely on the output from PowerPath.  Back in RedHat 4.5 days we actually had some issues where running sympd list would panic the box, so system admin got kind of skittish after that. We since have moved on to RHEL 6.x so it has not been an issue. Just like you i like the inq utility, drop it, make it executable (on *nix) and go ..no need to mess around with solutions enabler.

Thanks guys I will try the inq .

For people happy with UNIX, grep can be handy for you.

root@rox-dev $ symdev -sid 1234 list -identifier device_name -devs 1900:1a0a | grep vSphere

Symmetrix ID: 000192601234

                                  Device

-------------------------------------------------------------------------------

Sym  Config          Attr Device Name

---- --------------- ---- -----------------------------------------------------

19F2 TDEV            (M)  vSphere_Prod_Cluster

19FA TDEV            (M)  vSphere_Prod_Cluster

1A02 TDEV            (M)  vSphere_Prod_Cluster

1A0A TDEV            (M)  vSphere_Prod_Cluster

I use this output to check the devices out by name, before performing operations on them.

You can do the same with a simple FOR loop, that lists the identifiers for all devs in a storage group.

Thanks!

I have another question , some what related to this thread .

When a sys admin asks for a disk to be expanded , can we go ahead and expand any lun and then they can allocate it from their end ?

For e.g ,

Hostname : Kronos || D:150 GB || E:1TB||F:500 GB||

Luns created as per inital request

0EFC- 150 GB

0EFD - 1 TB

0EFE- 500 GB .

Now host needs an additional 500 Gigs on D ...

I can expand any dev  ? or I should expand 0EFC only ?

Is there any place where I can read and gain understaning such situations  . Its the most basic thing I should be knowing and I cant seem to get my head around this simple logic ....

Cheers

it's a one to one relationship (especially on windows with default volume manager). One symmetrix device represents one drive letter on the host side. One some windows systems this could be a mount point.  Does it help ?

If its a one to one relationship then I shoud be expanding the same tdev .

Its always good to check the how  the devs are mapped at the host end . I shouldn't be expanding any devs blindly

is what I get   .

Can multiple devs(meta)  be mapped to a single drive letter ? Let say I have two metaheads 0EFC&0EFE , can they be merged into a single drive .

Want to understand how it works in unix/windows .. I've read about primary/extened /dynamic disks .. 

Thanks Dynamox .