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

Re: [dm-devel] Multipath Usage Guide available at LVM wiki page



On Thu, 2008-01-24 at 18:14 -0800, malahal us ibm com wrote:
> Chandra Seetharaman [sekharan us ibm com] wrote:
> > On Thu, 2008-01-24 at 10:33 -0800, Pradipmaya Maharana wrote:
> > > Hi Chandra,
> > > 
> > > Thanks for the document, it surely is helpful.
> > > 
> > > I had a query (which I have asked in previous mails), the document
> > > does not talk about how to install SLES on multipath devices with LVM.
> > > Basically I tried installing with creating partitions (for both RHEL
> > > and SLES) and it worked fine; LVM worked fine for RHEL but for SLES it
> > > won't work.
> > > 
> > > - It won't let me mount-by device-id for LVM
> > > - specifying the volume label won't help.
> 
> Are you saying it won't help or it did NOT help?
> 
> > You could put your lvm on top of the multipathed dm-device. For which
> > you may have to disable lvm probing in the initrd phase. You can do this
> > by changing your /etc/lvm/lvm.conf and then running mkinitrd.
> > 
> > You have to change the filter line to be like this:
> > 
> > filter = [ "a|/dev/mapper.*|", "a|/dev/disk/by-name/.*|", "r|.*|" ]
> > 
> > Let me know if this helps.
> 
> multipath should be run before lvm (even in initrd), if so, how does
> this filter help?  LVM will not be able to claim plain devices in the
> presence of multipath devices anyway, right?

It can. I have seen it happen (in sles9 though). multipath opens the
(sd) device and later lvm scans the same device, it fails (as there is
already an open reference). 

The filter above helps in making lvm read the multipathed device instead
of the underlying scsi device, and hence the lvm scan succeeds.

> 
> --
> dm-devel mailing list
> dm-devel redhat com
> https://www.redhat.com/mailman/listinfo/dm-devel
-- 

----------------------------------------------------------------------
    Chandra Seetharaman               | Be careful what you choose....
              - sekharan us ibm com   |      .......you may get it.
----------------------------------------------------------------------



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