Help with multipath failover

Brian O'Mahony brian.omahony at curamsoftware.com
Thu Apr 8 04:56:10 CDT 2010


I tested the failover using iozone. There was no interruption when I unplugged a cable, but when I put that back in, waited about 20sec and unplugged the other the IOPS stopped. But I reckon that is just natural behavior. I probably didn’t allow long enough for the failover.

Thanks for all the help.

B

-----Original Message-----
From: Charles Riley [mailto:criley at erad.com] 
Sent: Tuesday, April 06, 2010 6:27 PM
To: Brian O'Mahony
Subject: Re: Help with multipath failover

To check if the IO is interrupted, start a process that creates io to and or from the multipathed partition.
There are all sorts of ways to do that.
If multipath is working the io will complete even if one of the paths is lost.

To flush the multipath configuration you first have to unmount any multipathed devices.

Charles Riley
eRAD, Inc.



----- "Brian O'Mahony" <brian.omahony at curamsoftware.com> wrote:

> Sorry for the delay in responding to this. I was away at the end of
> last week.
> 
> Anyways, to test I was running a ping from one terminal and then using
> the other terminal to run multipath -ll, and as far as I can see the
> actual multipathing is fine. However the machine isn't in the office
> and there is no-one on site till the morning.
> 
> What method would you suggest to measure if the I/O is interrupted? I
> would have thought if the ping was unsuccessful, then so to would be
> read/writes?
> 
> I also have two other quick questions:
> 
> 1. Where do you go about deleting multipath configuration? When I use
> multipath -F it tells me the map is in use, so I stopped the service
> and logged out of the node, but the error stays the same.
> 
> 2. Is there any tools you would recommend for monitoring read/write
> access to the SAN once it is in production? I just want to be able to
> monitor, preferably in real time, the read/write speeds, and whether
> there is bottlenecks - eg with the network to the SAN.
> 
> Thanks for all the help
> B
> 
> -----Original Message-----
> From: Shyam_Iyer at Dell.com [mailto:Shyam_Iyer at Dell.com] 
> Sent: Wednesday, March 31, 2010 7:51 PM
> To: Brian O'Mahony; criley at erad.com;
> linux-poweredge at lists.us.dell.com
> Subject: RE: Help with multipath failover
> 
> 
> 
> Well you are testing link failover and not I/O failover then .. (:
> 
> 
> I would ideally try to do I/O through one path yank out the cable and
> watch the I/O failover to the other path...
> That is probably what you want to ensure if you want to use
> dm-multipath..
> 
> Also, make sure you change the configuration for failback from
> "immediate" to "queue" so that you won't see immediate I/O failures
> and see the I/O safely failover to the other path.
> 
> Thanks,
> Shyam Iyer
> Linux Engineering.
> 
> -----Original Message-----
> From: linux-poweredge-bounces at dell.com on behalf of Brian O'Mahony
> Sent: Wed 3/31/2010 2:02 PM
> To: 'criley at erad.com'; linux-poweredge-Lists
> Subject: Re: Help with multipath failover
>  
> Ping to san ip and yanking a cable....
> ------------------------
> Sent from my BlackBerry Wireless Handheld
> 
> 
> ----- Original Message -----
> From: Charles Riley <criley at erad.com>
> To: Brian O'Mahony
> Sent: Wed Mar 31 18:34:12 2010
> Subject: Re: Help with multipath failover
> 
> How are you determining that it doesn't fail over?
> 
> Charles Riley
> eRAD, Inc.
> 
> 
> ----- "Brian O'Mahony" <brian.omahony at curamsoftware.com> wrote:
> 
> > Im connecting a PE2850 to a Equalogic SAN, using multipath. I have
> a
> > PCI NIC with one connection to the SAN and one from the onboard
> > controller too. The SAN is ona different network, and is at IP
> > 10.10.20.10.
> > 
> > 
> > 
> > Here is the output from the multipath -ll
> > 
> > [root at ccvobtest2850 ~]# multipath -ll
> > 
> > DubSanGrp01VolCC (36090a048d03f308d51b1f47c000030a6) dm-18
> > EQLOGIC,100E-00
> > 
> > [size=500G][features=1 queue_if_no_path][hwhandler=0][rw]
> > 
> > \_ round-robin 0 [prio=1][enabled]
> > 
> > \_ 3:0:0:0 sdc 8:32 [active][ready]
> > 
> > \_ round-robin 0 [prio=1][enabled]
> > 
> > \_ 4:0:0:0 sdd 8:48 [active][ready]
> > 
> > 
> > 
> > 
> > 
> > 
> > 
> > And here is the multipath.conf:
> > 
> > [root at ccvobtest2850 ~]# cat /etc/multipath.conf
> > 
> > ## Use user friendly names, instead of using WWIDs as names.
> > 
> > defaults {
> > 
> > user_friendly_names yes
> > 
> > }
> > 
> > blacklist {
> > 
> > # wwid 26353900f02796769
> > 
> > # devnode "^(ram|raw|loop|fd|md|dm-|sr|scd|st)[0-9]*"
> > 
> > # devnode "^hd[a-z]"
> > 
> > # devnode "^sd[a-b]"
> > 
> > }
> > 
> > multipaths {
> > 
> > multipath {
> > 
> > wwid 36090a048d03f308d51b1f47c000030a6
> > 
> > alias DubSanGrp01VolCC
> > 
> > }
> > 
> > }
> > 
> > devices {
> > 
> > device {
> > 
> > vendor "EQLOGIC"
> > 
> > product "100E-00"
> > 
> > path_grouping_policy failover
> > 
> > getuid_callout "/sbin/scsi_id -g -u -s /block/%n"
> > 
> > features "1 queue_if_no_path"
> > 
> > path_checker readsector0
> > 
> > failback immediate
> > 
> > path_selector "round-robin 0"
> > 
> > rr_min_io 10
> > 
> > rr_weight priorities
> > 
> > }
> > 
> > }
> > 
> > 
> > 
> > 
> > 
> > However when I unplug the cable to eth1, it doesn't failover. I am
> > testing with a ping to the SAN at 10.10.20.10. As Im pretty new to
> all
> > this, was wondering if anyone has any pointers on where I am going
> > wrong.
> > 
> > 
> > 
> > Thanks
> > 
> > Brian
> > 
> > 
> > 
> > The information in this email is confidential and may be legally
> > privileged.
> > It is intended solely for the addressee. Access to this email by
> > anyone else
> > is unauthorized. If you are not the intended recipient, any
> > disclosure,
> > copying, distribution or any action taken or omitted to be taken in
> > reliance
> > on it, is prohibited and may be unlawful. If you are not the
> intended
> > addressee please contact the sender and dispose of this e-mail.
> Thank
> > you. 
> > _______________________________________________
> > Linux-PowerEdge mailing list
> > Linux-PowerEdge at dell.com
> > https://lists.us.dell.com/mailman/listinfo/linux-poweredge
> > Please read the FAQ at http://lists.us.dell.com/faq
> 
> 
> The information in this email is confidential and may be legally
> privileged.
> It is intended solely for the addressee. Access to this email by
> anyone else
> is unauthorized. If you are not the intended recipient, any
> disclosure,
> copying, distribution or any action taken or omitted to be taken in
> reliance
> on it, is prohibited and may be unlawful. If you are not the intended
> addressee please contact the sender and dispose of this e-mail. Thank
> you.
> 
> _______________________________________________
> Linux-PowerEdge mailing list
> Linux-PowerEdge at dell.com
> https://lists.us.dell.com/mailman/listinfo/linux-poweredge
> Please read the FAQ at http://lists.us.dell.com/faq
> 
> 
> 
> The information in this email is confidential and may be legally
> privileged.
> It is intended solely for the addressee. Access to this email by
> anyone else
> is unauthorized. If you are not the intended recipient, any
> disclosure,
> copying, distribution or any action taken or omitted to be taken in
> reliance
> on it, is prohibited and may be unlawful. If you are not the intended
> addressee please contact the sender and dispose of this e-mail. Thank
> you.
> 
> 
> _______________________________________________
> Linux-PowerEdge mailing list
> Linux-PowerEdge at dell.com
> https://lists.us.dell.com/mailman/listinfo/linux-poweredge
> Please read the FAQ at http://lists.us.dell.com/faq


The information in this email is confidential and may be legally privileged.
It is intended solely for the addressee. Access to this email by anyone else
is unauthorized. If you are not the intended recipient, any disclosure,
copying, distribution or any action taken or omitted to be taken in reliance
on it, is prohibited and may be unlawful. If you are not the intended
addressee please contact the sender and dispose of this e-mail. Thank you.



More information about the Linux-PowerEdge mailing list