linux device mapper important questions

[bitou]

hi all,

i installed the device mapper on a redhat 5.4 (32 bits ) machine. unfortunately, the server was connected to the SAN switch utilizing two fiber channel connection from the HBA card. the controller is connect to the san switch and also to the enclosure. 

normally we have two paths before installed the multipathing software.

Before the installation of HP device mapper, the server saw 2 vdisks if one (Vdisk) was presented from Command View, and it is normal.

after the installation the 2 Vdisks becomes one Vdisk and also the situation is normal.

i have created a physial volume, vg, LV,....

pvreate /dev/mapper/mpath1

vgcreate vg01 /dev/mapper/mpath1

lvcreate -l 2 GB -n lvoltest

mkfs ext3...

mount .....

my problem is:

how to verify the multipathing is working properlly?

I have like result the following:

disconnect FC 2 cable from on port of HBA card: the server is still see the VG and Vdisk

disconnect FC1 and connect FC2: the VDisk was disappeared.

please i need the device mapper command to verify if the two paths were active active....

is it true the command multipath -l

thanks in advance

[Turgay Cavdar]

Hi ,
Can you please post the output of:
# multipath -l

# cat /etc/multipath.conf

[Steven E. Protter]

Shalom,

Clearly from your results multi pathing is not set up correctly OR more likely the disk is not properly presented to the second HBA card.

SEP

[bitou]

thanks for your responses :

the result

FC1 cable to port 1 on HBA connection, port 2 HBA is disconnet

of [root@linuxmulti ~]# multipath -ll
[size=2.0G][features=1 queue_if_no_path][hwhandler=0][rw]
\_ round-robin 0 [prio=50][active]
 \_ 0:0:1:1 sdb 8:16  [active][ready]
\_ round-robin 0 [prio=10][enabled]
 \_ 0:0:0:1 sda 8:0   [active][ready]

FC1 cable to port 2 on HBA connection, port 1 HBA is disconnet

of [root@linuxmulti ~]# multipath -ll
\[size=2.0G][features=1 queue_if_no_path][hwhandler=0][rw]
\_ round-robin 0 [prio=50][active]
 \_ 0:0:1:1 sdb 8:16  [active][ready]
\_ round-robin 0 [prio=10][enabled]
 \_ 0:0:0:1 sda 8:0   [active][ready]

FC2 cable to port 1 on HBA connection, port 2 HBA is disconnet

of [root@linuxmulti ~]# multipath -ll
[size=2.0G][features=1 queue_if_no_path][hwhandler=0][rw]
\_ round-robin 0 [prio=0][enabled]
 \_ 0:0:1:1 sdb 8:16  [failed][faulty]
\_ round-robin 0 [prio=0][enabled]
 \_ 0:0:0:1 sda 8:0   [failed][faulty]

FC2 cable to port 2 on HBA connection, port 1 HBA is disconnetof

[root@linuxmulti ~]# multipath -ll
[size=2.0G][features=1 queue_if_no_path][hwhandler=0][rw]
\_ round-robin 0 [prio=0][enabled]
 \_ 0:0:1:1 sdb 8:16  [failed][faulty]
\_ round-robin 0 [prio=0][enabled]

and also the vdisk is presented for both wwn on the HBA. it is normal the result of the command multipath.

[bitou]

any updates please :smileyhappy:

[Turgay Cavdar]

Hi,

From multipath output, what i can see is h-c-t-l address (0:0:1:1) part is always zero, this means you can only see devices from one hba port.  

[Alzhy]

Bitou,

1.) What storage SAN Array are you using

2.) Are both your HBAs zoned properly with your SAN array?

3. ) What version of HP Device Mapper and version of Linux RHEL?

If it is an EVA6000+, 8000+ - you should see 8 device paths for every vdisk if zoning is properly done.

WIthin /etc/multipath.conf -- you can also name your disk friendlier - say /dev/mapper/EVAdisk01 instead of /dev/mapper/mpath1.

Pls attach your /etc/multipath.conf and ask your SAN admin for the Zone entries for your Server,