[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

[dm-devel] RE: Reducing amount of logmessages openiscsi/multipath [with md3000i]




> -----Original Message-----
> He wants the RDAC hw handler and its path checker (rdac). The MD3000i
> is an
> LSI re-branded box and is the same as the IBM DS3300.
> 
> >
> > You also want to make sure that you are using the md3000i hw handler
> or
> > scsh_dh_module if you are not already. The dm-devel guys can help you
> out.
> 
> Here is the multipath.conf he should be using for MD3000i (or for the
> DS3300
> but will need to modify the vendor/model entry)
> 
>         # Note: The same as the IBM DS3300
>         #
>         device {
>                 vendor                  "DELL"
>                 product                 "MD3000i"
>                 product_blacklist       "Universal Xport"
>                 features                "1 queue_if_no_path"
>                 path_grouping_policy    group_by_prio
>                 hardware_handler        "1 rdac"
>                 path_checker            rdac
>                 prio                    "rdac"
>                 failback                immediate
>         }
> 
> 

This fixed the problem, although a simple flush was not enough to activate
the changes, so it took a bit longer to get the result I wanted. In the end
I saw that it was still using the direction path_checker and noticed that
you cannot change the 'path_checker' on the fly with just a multipath -F /
-v 2, but that a reload of the modules was needed.

$ multipath -ll
webdata (36001ec9000d163110000067f484e260c) dm-0 DELL    ,MD3000i
[size=2.0T][features=0][hwhandler=1 rdac]
\_ round-robin 0 [prio=1000][active]
 \_ 0:0:0:0 sda 8:0   [active][ready]
\_ round-robin 0 [prio=1][enabled]
 \_ 1:0:0:0 sdb 8:16  [active][ready]
\_ round-robin 0 [prio=0][enabled]
 \_ 2:0:0:0 sdc 8:32  [active][ghost]
 \_ 3:0:0:0 sdd 8:48  [active][ghost]

That is the output of multipath now, and no more spam in my syslog or dmesg.

(The different priorities you see are because the second path (with a prio
of 1) is a longer path through 1 more switch, so I'd rather not give it a
high priority, and if the '1000' path fails, it just uses the '1' path.)

A big thanks to everyone who directed me on the right path ;)

-kees


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]