Resizing disks - ASM mulpath

[satish]

Hi Erman,


we are using RAC and maultipathing
Our version 12.1.0.2

we have a planned activity of Resizing Of ASM Disks.Our storage administrator requested us to give the correct disk names,so that he can cross check before extending.

To get correct disks we looked into ASM as follows


select g.name
     , d.path
     , d.os_mb
     , d.total_mb
from v$asm_diskgroup g
   , v$asm_disk      d
where g.group_number = d.group_number
and   g.name = 'FRADG'
/

NAME       PATH                 OS_MB   TOTAL_MB
---------- --------------- ---------- ----------
FRADG      ORCL:ARCHDISK1      102398     102396
FRADG      ORCL:ARCHDISK5      102398     102396
FRADG      ORCL:ARCHDISK4      102398     102396
FRADG      ORCL:ARCHDISK3      102398     102396
FRADG      ORCL:ARCHDISK2      102398     102396

SQL>

[gridprod@node1dev]$ ls -l /dev/oracleasm/disks/ARCH*
brw-rw---- 1 gridprod asmadmin 253, 69 Apr 23  2018 /dev/oracleasm/disks/ARCHDISK1
brw-rw---- 1 gridprod asmadmin 253, 76 Apr 23  2018 /dev/oracleasm/disks/ARCHDISK2
brw-rw---- 1 gridprod asmadmin 253, 75 Apr 23  2018 /dev/oracleasm/disks/ARCHDISK3
brw-rw---- 1 gridprod asmadmin 253, 77 Apr 23  2018 /dev/oracleasm/disks/ARCHDISK4
brw-rw---- 1 gridprod asmadmin 253, 78 Apr 23  2018 /dev/oracleasm/disks/ARCHDISK5
[gridprod@node1dev]$


This gives is the major and minor numbers for the disks – major number is 253 and minor numbers are 69,76,75,77,78

Then comparing these numbers against the devices listed in /dev/mapper and got the below output which is not clear to me

[gridprod@erpproddb1 dev]$ ls -l /dev/mapper/*ARCH*
lrwxrwxrwx 1 root root 8 Jun 22  2018 /dev/mapper/PRD-ARCH01 -> ../dm-19
lrwxrwxrwx 1 root root 8 Apr 23  2018 /dev/mapper/PRD-ARCH01p1 -> ../dm-69
lrwxrwxrwx 1 root root 8 Jun 22  2018 /dev/mapper/PRD-ARCH02 -> ../dm-17
lrwxrwxrwx 1 root root 8 Apr 23  2018 /dev/mapper/PRD-ARCH02p1 -> ../dm-76
lrwxrwxrwx 1 root root 8 Jun 22  2018 /dev/mapper/PRD-ARCH03 -> ../dm-20
lrwxrwxrwx 1 root root 8 Apr 23  2018 /dev/mapper/PRD-ARCH03p1 -> ../dm-75
lrwxrwxrwx 1 root root 8 Jun 22  2018 /dev/mapper/PRD-ARCH04 -> ../dm-21
lrwxrwxrwx 1 root root 8 Apr 23  2018 /dev/mapper/PRD-ARCH04p1 -> ../dm-77
lrwxrwxrwx 1 root root 8 Jun 22  2018 /dev/mapper/PRD-ARCH05 -> ../dm-22
lrwxrwxrwx 1 root root 8 Apr 23  2018 /dev/mapper/PRD-ARCH05p1 -> ../dm-78
[gridprod@node1dev]$


could you please help in getting the disknames to provide them to administrator

Thank you for the support

[ErmanArslansOracleBlog]

check this blog post -> https://blogs.oracle.com/alejandrovargas/mapping-asm-disks-to-physical-devices

[satish]

Dear Erman,

This is our multipath -ii output

PRD-ARCH01 (360002ac0000000000000006b0001c9fc) dm-19 3PARdata,VV
size=100G features='1 queue_if_no_path' hwhandler='1 alua' wp=rw
`-+- policy='round-robin 0' prio=50 status=active
  |- 2:0:0:24 sdo  8:224   active ready running
  |- 1:0:1:24 sdec 128:64  active ready running
  |- 2:0:1:24 sdbv 68:144  active ready running
  `- 1:0:0:24 sdgj 131:240 active ready running


PRD-ARCH02 (360002ac0000000000000006c0001c9fc) dm-17 3PARdata,VV
size=100G features='1 queue_if_no_path' hwhandler='1 alua' wp=rw
`-+- policy='round-robin 0' prio=50 status=active
  |- 2:0:0:25 sdp  8:240   active ready running
  |- 1:0:1:25 sded 128:80  active ready running
  |- 2:0:1:25 sdbw 68:160  active ready running
  `- 1:0:0:25 sdgk 132:0   active ready running


PRD-ARCH03 (360002ac0000000000000006d0001c9fc) dm-20 3PARdata,VV
size=100G features='1 queue_if_no_path' hwhandler='1 alua' wp=rw
`-+- policy='round-robin 0' prio=50 status=active
  |- 2:0:0:26 sdq  65:0    active ready running
  |- 1:0:1:26 sdee 128:96  active ready running
  |- 2:0:1:26 sdbx 68:176  active ready running
  `- 1:0:0:26 sdgl 132:16  active ready running


PRD-ARCH04 (360002ac0000000000000006e0001c9fc) dm-21 3PARdata,VV
size=100G features='1 queue_if_no_path' hwhandler='1 alua' wp=rw
`-+- policy='round-robin 0' prio=50 status=active
  |- 2:0:0:27 sdr  65:16   active ready running
  |- 1:0:1:27 sdef 128:112 active ready running
  |- 2:0:1:27 sdby 68:192  active ready running
  `- 1:0:0:27 sdgm 132:32  active ready running


PRD-ARCH05 (360002ac0000000000000006f0001c9fc) dm-22 3PARdata,VV
size=100G features='1 queue_if_no_path' hwhandler='1 alua' wp=rw
`-+- policy='round-robin 0' prio=50 status=active
  |- 2:0:0:28 sds  65:32   active ready running
  |- 1:0:1:28 sdeg 128:128 active ready running
  |- 2:0:1:28 sdbz 68:208  active ready running
  `- 1:0:0:28 sdgn 132:48  active ready running

[satish]

we are using device mapper

[ErmanArslansOracleBlog]

Okay good.
In multipath output, you already see the device names :

for ex here ->

PRD-ARCH01 (360002ac0000000000000006b0001c9fc) dm-19 3PARdata,VV
size=100G features='1 queue_if_no_path' hwhandler='1 alua' wp=rw
`-+- policy='round-robin 0' prio=50 status=active
  |- 2:0:0:24 sdo  8:224   active ready running
  |- 1:0:1:24 sdec 128:64  active ready running
  |- 2:0:1:24 sdbv 68:144  active ready running
  `- 1:0:0:24 sdgj 131:240 active ready running

This means your multipath device named PRD-ARCH01 consisting of 4 100G disks, which are sdo,sdec,sdby,sdgj..
There are also scsi ids there ->  for ex: sdo ->  2:0:0:24 .
This means Your sdo disk is coming from 2nd scsi host, Channel :0 Id: 02 Lun: 24..

Analyze the whole output of multipath like this..

Tell them to your SAN admin.

What else do you need?

[satish]

Dear Erman,

Thanks for the update.We are on RHEL 7,so we have modified -d to the script to give the output

[root@erpproddb1 ~]#  /usr/sbin/oracleasm querydisk DATADISK1
Disk "DATADISK1" is a valid ASM disk
[root@erpproddb1 ~]#


It didnt give the major and minor numbers as shown above.so we added -d and then we are able to get them

[root@erpproddb1 ~]# /usr/sbin/oracleasm querydisk -d DATADISK1
Disk "DATADISK1" is a valid ASM disk on device [253,81]
[root@erpproddb1 ~]#

Thanks for all the support

[satish]

Thanks for the update.

Gone through the updates,need one clarification

1)Are the below disks OR paths -
sdo,sdec,sdby,sdgj.

2)dm-19 is not physical disk right?

PRD-ARCH01 (360002ac0000000000000006b0001c9fc) dm-19 3PARdata,VV
size=100G features='1 queue_if_no_path' hwhandler='1 alua' wp=rw
`-+- policy='round-robin 0' prio=50 status=active
  |- 2:0:0:24 sdo  8:224   active ready running
  |- 1:0:1:24 sdec 128:64  active ready running
  |- 2:0:1:24 sdbv 68:144  active ready running
  `- 1:0:0:24 sdgj 131:240 active ready runni

Thanks,
Satish

[ErmanArslansOracleBlog]

your disks are the ones named with sd prefix..
multipath device (for ex: dm-19) is a logical one, which directs the I/O to those disks.