[ofa-general] SRP Failover

[PN Lai]

I use RHEL, it works very fine. Thanks.

 

I have another question.

I tried with a normal server (without RAID controller) to simulate the
storage and it cannot be recognized by multipath.

Does it mean that I can't use a normal server (without RAID controller) to
simulate the storage?

Since the WWID used in multipath seems to be generated by the RAID
controller.

 

Thanks again.

PN

 

From: Scott Weitzenkamp (sweitzen) [mailto:sweitzen at cisco.com] 
Sent: Tuesday, June 26, 2007 11:58 PM
To: PN Lai; general at lists.openfabrics.org; Scott Weitzenkamp (sweitzen)
Subject: RE: [ofa-general] SRP Failover

 

You need to configure Device Mapper Multipath or some other multipathing
software to get HA.  What OS are you running?

 

Steps for RHEL are:

 

1) Edit /etc/multipath.conf and comment out devnode_blacklist (RHEL4) or
blacklist (RHEL5) entry.

2) Run "chkconfig multipathd on".

3) Reboot.

4) After reboot, /dev/mapper should be populated with mutipath block device
entries.

5) You can run "multipath -l" to view the multipath status.

 

Steps for SLES10 are similar:

 

1) Run "chkconfig boot.multipath on".

2) Run "chkconfig multipathd on".

3) Reboot.

4) After reboot, /dev/mapper should be populated with mutipath block device
entries.

5) You can run "multipath -l" to view the multipath status.

 

You use the /dev/mapper block devices, not /dev/sd* block devices.

 

Scott Weitzenkamp

SQA and Release Manager

Server Virtualization Business Unit

Cisco Systems

 

 


  _____  


From: general-bounces at lists.openfabrics.org
[mailto:general-bounces at lists.openfabrics.org] On Behalf Of PN Lai
Sent: Tuesday, June 26, 2007 2:48 AM
To: general at lists.openfabrics.org
Subject: [ofa-general] SRP Failover

Hi all,

 

I'm testing the SRP HA functions, but I have some questions.

I use 2 IB cables to connect the initiator and 1 IB cables to connect to the
storage.

 

I installed the OFED-1.2, enable the "SRP_LOAD=yes" and "SRPHA_ENABLE=yes"
in openib.conf.

After reboot, it discovers 2 targets /dev/sdbX and /dev/sdcX. 

 

However, I check the /var/log/srp_daemon.log, it shows:

..

26/05/07 17:42:57 : bad MAD status (110) from lid 257

26/05/07 17:43:30 : No response to inform info registration

26/05/07 17:43:30 : Fail to register to traps, maybe there is no opensm
running on fabric

..

 

But the opensm is running in both machines. I don't know whether it is
normal, or should it only discover a single target?

 

Now, my question is that if I mount the /dev/sdbX and write data to it, and
then remove 1 of the initiator cable, how the /dev/sdcX will replace the
/dev/sdbX so that I can continue to write the data?

 

Do I need to configure some extra files?

 

Thanks for reply.

 

PN

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openfabrics.org/pipermail/general/attachments/20070628/96476540/attachment.html>