[dm-devel] Re: multipath's "-l" option behaves differently from "-ll" - for blacklisted devices
Amit K. Arora
aarora at linux.vnet.ibm.com
Fri Jan 19 08:34:11 UTC 2007
Can someone please comment on this behavior ? Please let me know if the
question is not clear and I need to rephrase it in different words.
Thanks!
--
Regards,
Amit Arora
> Hello,
>
> A little bit of background first...
>
> If we have blacklisted devices in /etc/multipath.conf, "multipath -l"
> output looks like this (please note "#"s below) :
> # multipath -l
> mpath2 () dm-4 ,
> [size=15G][features=0][hwhandler=0]
> \_ round-robin 0 [prio=0][active]
> \_ #:#:#:# - #:# [active][undef]
> \_ #:#:#:# - #:# [active][undef]
> mpath1 () dm-1 ,
> [size=15G][features=0][hwhandler=0]
> \_ round-robin 0 [prio=0][active]
> \_ #:#:#:# - #:# [active][undef]
> \_ #:#:#:# - #:# [active][undef]
> mpath3 () dm-7 ,
> [size=15G][features=0][hwhandler=0]
> \_ round-robin 0 [prio=0][active]
> \_ #:#:#:# - #:# [active][undef]
> \_ #:#:#:# - #:# [active][undef]
>
>
> And, for these blacklisted devices, "multipath -ll" gives full output
> (i.e. no "#"s in the result) :
> # multipath -ll
> mpath2 () dm-4 IBM,1724-100 FAStT
> [size=15G][features=0][hwhandler=0]
> \_ round-robin 0 [prio=2][active]
> \_ 1:0:0:2 sdd 8:48 [active][ready]
> \_ 1:0:1:2 sdh 8:112 [active][ready]
> mpath1 () dm-1 IBM,1724-100 FAStT
> [size=15G][features=0][hwhandler=0]
> \_ round-robin 0 [prio=2][active]
> \_ 1:0:0:1 sdc 8:32 [active][ready]
> \_ 1:0:1:1 sdg 8:96 [active][ready]
> mpath3 () dm-7 IBM,1724-100 FAStT
> [size=15G][features=0][hwhandler=0]
> \_ round-robin 0 [prio=2][active]
> \_ 1:0:0:3 sde 8:64 [active][ready]
> \_ 1:0:1:3 sdi 8:128 [active][ready]
>
> >From the source code we can see that in both cases (-l and -ll),
> path_discovery() and get_dm_mpvec() get called. The only difference is
> that -ll will result in a call to update_paths() in get_dm_mpvec(),
> after multipath maps have been disassembled/discovered.
>
> The question is that, is this behavior intended, a bug or a "feature" as
> a result of the design of the tool ? And, does it make sense to keep the
> behavior identical for -l and -ll options (of multipath tool) incase of
> blacklisted devices ?
>
> Thanks!
> --
> Regards,
> Amit Arora
More information about the dm-devel
mailing list