12:17 a.m.
Hi,
if you change the IP the procedure basically is this one ("Changing a
Server NID"):
http://wiki.old.lustre.org/manual/LustreManual20_HTML/LustreMaintenance.h...
However, you wrote that your setup is already a failover configuration,
in which both targets (MDT and MGT) by default are on the same host, and
both targets can fail over to a standby server. So, I assume both
targets are connected by a storage network, are probably mapped via
muitipath, and you have some ha mechanism like heartbeat and stonith or
something, which triggers the failover, right?
http://wiki.old.lustre.org/manual/LustreManual20_HTML/ConfiguringFailover...
If that's the case, I think the only step which is really needed is that
you change the configuration of the ha, so that it mounts the mgt on the
second server by default.
You also can go through the procedure of changing the server nid and
swap the two MGS-nids. Note that there is the --print and the --dryrun
option for tunefs.lustre, so you can check your current settings and try
out the effect of the combination of several parameters.
On the other hand I don't believe that splitting the failover of MGT and
MDT would change much in the timing how long this takes. This should
also be somewhere in the ha configuration, and recovery might also play
a role here. A look at the syslog of the servers could indicate what's
happening during failover, and what's the reason that it takes that
long. But anyhow, failover always takes some time. The failover host
must ensure that the other one has not mounted the targets anymore, by
trying to ping, log in, check mtab, trigger umount, and/or shoot it down
(each action with some reasonable timeout), then it must mount, and
after that recovery would probably start and clients start to notice
that the failover has happened and connect to the failover host. 20
minutes is not that incredibly long for all this.
Martin
On 07/23/2015 02:14 PM, Indivar Nair wrote:
Hi Andreas,
Thanks for the input.
I checked the document and found this topic -
https://build.hpdd.intel.com/job/lustre-manual/lastSuccessfulBuild/artifa...
I have MGT and MDT on separate devices.
So is there anyway to reconfigure the MGT to use the new default and
failover IP without formatting it?
In any case, how do I then tell the MDTs and OSTs that the MGT's default
and failover IPs have changed?
I cant find these in the document.
Thanks and Regards,
Indivar Nair
On Wed, Jul 22, 2015 at 5:25 AM, Dilger, Andreas <andreas.dilger(a)intel.com>
wrote:
> I believe this is described in the Lustre Manual, but the basic process to
> split a combined MDS+MGS into a separate MGS is to format a new MGS device,
> then copy all the files from CONFIGS on the old combined MDT+MGT device
> into the new MGS. See the manual for full details.
>
> Cheers, Andreas
>
> On Jul 21, 2015, at 01:27, Indivar Nair <indivar.nair(a)techterra.in<mailto:
> indivar.nair(a)techterra.in>> wrote:
>
> Hi ...,
>
> Currently, Failover and Recovery takes a very long long time in our setup;
> almost 20 Minutes. We would like to make it as fast as possible.
>
> I have two queries regarding this -
>
> 1.
> ===================================================
> The MGS and MDT are on the same host.
>
> We do however have a passive stand-by server for the MGS/MDT server, which
> only mounts these partitions in case of a failure.
>
> Current Setup
> Server A: MGS+MDT
> Server B: Failover MGS+MDT
>
> I was wondering whether I can now move the MGS or MDT Partition to the
> standby server (so that imperative recovery works properly) -
>
> New Setup
> Server A: MDT & Failover MGS
> Server B: MGS & Failover MDT
> OR
> Server A: MGS & Failover MDT
> Server B: MDT & Failover MGS
>
> i.e.
> Can I separate the MDT and MGS partitions on to different machines without
> formatting or reinstalling Lustre?
> ===================================================
>
> 2.
> ===================================================
> This storage is used by around 150 Workstations and 150 Compute (Render)
> Nodes.
>
> Out of these 150 workstations, around 30 - 40 are MS Windows. The MS
> Windows clients access the storage through a 2-node Samba Gateway Cluster.
>
> The Gateway Nodes are connected to the storage through a QDR Infiniband
> Network.
>
> We were thinking of adding NFS Service to the Samba Gateway nodes, and
> reconfiguring the Linux clients to connect via this gateway.
>
> This will bring down the direct Lustre Clients to just 2 nodes.
> So, will having only 2 clients improve the failover-recovery time?
> ===================================================
>
> Is there anything else we can do to speed up recovery?
>
> Regards,
>
>
> Indivar Nair
>