Statistic Lun Show issue

[tomekp]

Hello,

I've got really strange issue. I have two LUNs in single igroup.I can check the stastistics for one LUN, but not for the other one. Both LUNs are online, mapped, created in the same way, even at around the same time. Example:

cluster::> lun mapping show -vserver SVM01 -igroup my_igroup
Vserver Path Igroup LUN ID Protocol
---------- ---------------------------------------- ------- ------ --------
SVM01 /vol/lun001/lun001 my_igroup 0 fcp
SVM01 /vol/lun002/lun002 my_igroup 1 fcp
2 entries were displayed.

# let's check the stats for lun001:

cluster::> statistics lun show -interval 5 -iterations 5 -vserver SVM01 -lun /vol/lun001/lun001
Error: command failed: There are no entries matching your query.

# let's check the stats for lun002:

cluster::> statistics lun show -interval 5 -iterations 5 -vserver SVM01 -lun /vol/lun002/lun002
cluster : 10/5/2016 03:48:27
*Total Read Write Other Read Write Latency
Lun Vserver Ops Ops Ops Ops (Bps) (Bps) (ms)
------------------------------ ----------- ------ ---- ----- ----- ----- ----- -------
/vol/lun002/lun002 SVM01 0 0 0 0 0 0 -
[...]

If you can give me any clue why it doesn't work I would relaly appriciate it. The only "idea" I have is:

Node Hosting the LUN for lun001 is different than for lun002. Is there any setting on the Node level I can check?

cluster::> version
NetApp Release 8.3.1: Mon Aug 31 12:42:08 UTC 2015

Thanks in advance.

[tomekp]

tx for proposal but I cannot call you right now. And I would prefer to have a written answer for future readers sake 🙂

[Olalowo]

Kindly progress to precisely advise the part of my inputs that you find difficult to comprehend. I am on seat now, and I shall revert upon your feedback.

[Olalowo]

Hi Question-asker,

To put a total stop to this your problem I succintly recommend these following steps:

1. Separate out the LUNs.This will ensure they do not have the same LUN id #.

2. /vol/lun001/lun001 & /vol/lun002/lun002 are the 2 LUNs under consideration. With this, i strongly recommend you take out the second LUN(/vol/lun002/lun002).

3. Use this command

a. To confirm that both LUNs are actually & unequivocally online, use the command:

cluster::> lun show -m

The output should show you the 2 LUNS(/vol/lun001/lun001 & /vol/lun002/lun002) and also indicate the igroup to which they are connected with.You output should look like this:

LUN path                            Mapped to          LUN ID  Protocol
-----------------------------------------------------------------------
/vol/lun001/lun001 name_dgrp1        0     iSCSI
/vol/lun002/lun002 name_dgrp1        0     iSCSI

NB: I put name_dgrp1 because I do not know the name of your igroup. It is is not contained in your given info.

4.If the igroup is same for both, then you should use the command below to offline the 2nd LUN:

cluster::> lun offline /vol/lun002/lun002

5.Next you should now proceed to create a brand new volume for the second LUN, with the command,

vol create lunNEW -s volume hosting-aggr-name size [g]

NB:Say you are creating a new LUN with the name lunNEW and you want to only make it directly contained in a volume & not in a qtree, with  size 2g, to be created in an aggregate called A_SA_04. The complete command will be:

ssh phc-na-p005-s1 vol create lunNEW -s volume A_SA_04 2g

6.Afterwards, you progress the created volume to become a LUN with the command:

lun create -s 2g -t windows [ -o noreserve ] [ -e space_alloc ] /vol/lunNEW/lun0

7. Great you have now got your second LUN created in a brand new volume called lunNEW.

8.Finally map the LUN to your required igroup , i.e.

lun map [ -f ] /vol/lunNEW/lun0 <initiator_group> [ <lun_id> ]

AND THEN YOU CAN COMFORTABLY CHECK ON ANY LUN STATISTICS AT ANY TIME T USING YOUR COMMAN(the command you

statistics lun show -interval 5 -iterations 5 -vserver SVM01 -lun /vol/lunNEW/lun0

QED

NB:Should you still have any other problem, directly reach me at sammarch2016@gmail.com)

[tomekp]

Olalowo thanks for your output - however I don't understand what do you mean.

Why should I create new LUN and remove my old one? It just doesn't make sense to me. Currently I've got both LUNs on seperate volumes, mapped to the same igroup. Both are used in production so I cannot just offline a lun, and create a new one. Even if I could - I don't see what would be different to my current configuration?

[Olalowo]

Call me now on my deskphone so we have an online chat realtime.

Waiting now.

[tomekp]

Small update:

the same command later on worked, and a little bit later it again stopped working printing

Error: command failed: There are no entries matching your query

. Is this an known-bug?