[ewg] OFED 1.2-beta1 SRP multipath test results

[Chieng Etta]

Yes, I was running IO at that time.  After I removing the IB cable, the IO
stopped right away.  It took about 2~3 minutes until the first path failed
then the 2nd path picked up the IO.

Thanks,
Etta

-----Original Message-----
From: Scott Weitzenkamp (sweitzen) [mailto:sweitzen at cisco.com] 
Sent: Monday, April 09, 2007 5:14 PM
To: Chieng Etta; Ishai Rabinovitz
Cc: openfabrics-ewg at openib.org
Subject: RE: [ewg] OFED 1.2-beta1 SRP multipath test results

Etta, I'm curious, are you running traffic when you failover?  If so,
how long does it take for the traffic to recover after failover?

Scott Weitzenkamp
SQA and Release Manager
Server Virtualization Business Unit
Cisco Systems
 

> -----Original Message-----
> From: ewg-bounces at lists.openfabrics.org 
> [mailto:ewg-bounces at lists.openfabrics.org] On Behalf Of Chieng Etta
> Sent: Monday, April 09, 2007 3:04 PM
> To: 'Ishai Rabinovitz'
> Cc: openfabrics-ewg at openib.org
> Subject: RE: [ewg] OFED 1.2-beta1 SRP multipath test results
> 
> Hi Ishai,
> 
> Hi Ishai,
> 
> I have attached the log file on sles10 x86 and the steps to 
> test failover
> and failback.
> 
> I also tested on a sles10 x86_64 system today. The multipath 
> failover and
> failback were working fine.  However, after recovering the 
> failed path, the
> log file showed "remaining active paths: 1".  Is this correct?  
> I did the same steps on RedHat and the log message was 
> "remaining active
> paths: 2".
> 
> Thanks,
> Etta
> 
> -----Original Message-----
> From: Ishai Rabinovitz [mailto:ishai at dev.mellanox.co.il] 
> Sent: Sunday, April 08, 2007 12:55 AM
> To: Chieng Etta
> Cc: openfabrics-ewg at openib.org
> Subject: Re: [ewg] OFED 1.2-beta1 SRP multipath test results
> 
> Hi Chieng,
> 
> Thanks for the tests.
> In my tests the fail back works.
> Could you please describe the way you checked the fail back.
> Will it be possible to send me the /var/log/messages file.
> 
> Thanks
> Ishai
> 
> Chieng Etta wrote:
> > All,
> > 
> >  
> > 
> > SFW has tested SRP multipath functionality with 
> OFED-1.2-beta1 and here 
> > is the system information and results.
> > 
> >  
> > 
> > Initiator: rh4u4 x86_64 and sles10 x86, OFED-1.2-beta1
> > 
> > Target: rh4u3 x86_64, IBGD-1.8.3-rh16, Areca-1220 controller
> > 
> >  
> > 
> > Failover works on rh4u4 and sles10 with either 2 IB 
> connections from 
> > initiator or 2 IB connections from target.
> > 
> > However, failback only works on rh4u4 and not sles10.  Forcing path 
> > failback by using multipath command also does not work.
> > 
> >  
> > 
> > test6:~ # multipath -ll
> > 
> > mpath343 (SAreca_ARC-1220-VOL#00_54047306) ,
> > 
> > [size=1043G][features=0][hwhandler=0]
> > 
> > \_ round-robin 0 [prio=0][enabled]
> > 
> >  \_ #:#:#:# -   #:#   [failed][faulty]
> > 
> > \_ round-robin 0 [prio=0][active]
> > 
> >  \_ 2:0:0:0 sda 8:0   [active][ready]
> > 
> >  
> > 
> > test6:~ # multipath
> > 
> > : checker msg is "readsector0 checker reports path is down"
> > 
> > mpath345: failed to access path
> > 
> > SAreca_ARC-1220-VOL#00_54047306: rename mpath343 to mpath345
> > 
> > rename: mpath345 (SAreca_ARC-1220-VOL#00_54047306) 
> Areca,ARC-1220-VOL#00
> > 
> > [size=1043G][features=0][hwhandler=0]
> > 
> > \_ round-robin 0 [prio=0][undef]
> > 
> >  \_ 2:0:0:0 sda 8:0   [active][ready]
> > 
> > \_ round-robin 0 [prio=0][undef]
> > 
> >  \_ 7:0:0:0 sdc 8:32  [undef][ready]
> > 
> >  
> > 
> > test6:~ # multipath -ll
> > 
> > mpath345 (SAreca_ARC-1220-VOL#00_54047306) ,
> > 
> > [size=1043G][features=0][hwhandler=0]
> > 
> > \_ round-robin 0 [prio=0][enabled]
> > 
> >  \_ #:#:#:# -   #:#   [failed][faulty]
> > 
> > \_ round-robin 0 [prio=0][active]
> > 
> >  \_ 2:0:0:0 sda 8:0   [active][ready]
> > 
> >  
> > 
> >  
> > 
> > Etta Chieng
> > 
> > System Fabric Works
> > 
> > http://www.systemfabricworks.com/
> > 
> >  
> > 
> > 
> > 
> --------------------------------------------------------------
> ----------
> > 
> > _______________________________________________
> > ewg mailing list
> > ewg at lists.openfabrics.org
> > http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
>