I've seen a few cases where the client fails to pull updated configuration data. A reboot of that client should fix it.

-cf

On 03/12/2013 10:43 AM, Stu Midgley wrote:

unfortunately, that didn't work… I shutdown all clients, unmounted the mds and ost's and tuner.lustre --writeconf on all the lustre servers.

I then remounted the mds, then the ost's and everything appears to be just fine (ie. all the mounting worked etc). BUT on a client

OST000e : inactive device
OST000f : inactive device

and on the oss

2 UP obdfilter l1-OST000e l1-OST000e_UUID 5
3 UP obdfilter l1-OST000f l1-OST000f_UUID 5


and on the mds

19 UP osc l1-OST000e-osc-MDT0000 l1-MDT0000-mdtlov_UUID 5
20 UP osc l1-OST000f-osc-MDT0000 l1-MDT0000-mdtlov_UUID 5

On Wed, Mar 13, 2013 at 12:19 AM, Colin Faber <cfaber@gmail.com <mailto:cfaber@gmail.com>> wrote:

    Yeah basically it registered and now the MGS expects it. You're
    going to have to writeconf everything to get rid of that entry.

    -cf


    On 03/12/2013 10:15 AM, Stu Midgley wrote:

        Evening

        A mistake was made during the addition of 1 oss (2 ost's) to
        our online lustre file system.

        The ost was formatted with

        mkfs.lustre --ost --mgsnode=172.29.0.251 --fsname=l1 --index
        14 --mkfsoptions="-T largefile -m 0.5 -E
        stride=16,stripe-width=160" /dev/sdb


        and then mounted... and then quickly unmounted... and then
        reformatted with the same command and then attempted to be
        remounted...

        We now have a hung file system.

        on the mds I see

        Mar 12 13:38:09 mds kernel: LustreError: 140-5: Server
        l1-OST000e requested index 14, but that index is already in
        use. Use --writeconf to force

        now what should I do? Bring everything down, writeconf all the
        servers and bring everything back up? reformat the ost? lfs df
        -h on a client is showing

        OST000e : inactive device

        Thanks.


        --         Dr Stuart Midgley
        sdm900@sdm900.com <mailto:sdm900@sdm900.com>

        <mailto:sdm900@sdm900.com <mailto:sdm900@sdm900.com>>


        _______________________________________________
        HPDD-discuss mailing list
        HPDD-discuss@lists.01.org <mailto:HPDD-discuss@lists.01.org>

        https://lists.01.org/mailman/listinfo/hpdd-discuss





--
Dr Stuart Midgley
sdm900@sdm900.com <mailto:sdm900@sdm900.com>


_______________________________________________
HPDD-discuss mailing list
HPDD-discuss@lists.01.org
https://lists.01.org/mailman/listinfo/hpdd-discuss