4:59 a.m.
>>>> "TR" == Thomas Roth <t.roth(a)gsi.de>
writes:
Hi,
as far as I know, mgs failover with ZFS is still not implemented :( I
ran into the same problem. Didn't get around to fix this yet, but its on
my ToDo. If someone else is planning to do this, please let me know, so
I don't waste my time. See also https://jira.hpdd.intel.com/browse/LU-4334
Roland
-------
http://www.q-leap.com / http://qlustar.com
--- HPC / Storage / Cloud Linux Cluster OS ---
TR> Hi all, the specification of several '--mgsnode=' failover NIDS
TR> in OST formatting does not work in my test case. Using ':' to
TR> separate the NIDs works.
TR> Seems for the first time I have a failover pair of MDSes with
TR> Lustre 2.x. I formatted with:
TR> MDS:~# mkfs.lustre --mgs --mdt --fsname=testfs --index=0
TR> --servicenode=10.20.0.2@o2ib0 --servicenode=10.20.0.3@o2ib0
TR> /dev/md0
TR> No trouble mounting (the MGS/MDT is on 10.20.0.2).
TR> My test-OSS is a single box, so I did
TR> OSS:~# mkfs.lustre --reformat --ost --backfstype=zfs
TR> --fsname=testfs --index=0 -param ost.quota_type=ug
TR> --mgsnode=10.20.0.2@o2ib --mgsnode=10.20.0.3@o2ib lpool-oss/ost1
TR> raidz2 dm-1 dm-2 dm-3 dm-4 ...
TR> Permanent disk data:
TR> Target: testfs:OST0000 Index: 0 Lustre FS: testfs Mount type:
TR> zfs Flags: 0x62
TR> (OST first_time update )
TR> Persistent mount opts: Parameters: ost.quota_type=ug
TR> mgsnode=10.20.0.2@o2ib mgsnode=10.20.0.3@o2ib
TR> mkfs_cmd = zfs create -o canmount=off -o xattr=sa lpool-oss/ost0
TR> Writing lpool-oss/ost0 properties
TR> lustre:version=1 lustre:flags=98 lustre:index=0
TR> lustre:fsname=testfs lustre:svname=testfs:OST0000
TR> lustre:ost.quota_type=ug lustre:mgsnode=10.20.0.2@o2ib
TR> lustre:mgsnode=10.20.0.3@o2ib
TR> This didn't mount, claiming the MGS on 10.20.0.3@o2ib wasn't up
TR> - which is correct. I checked with tunefs.lustre:
TR> OSS:~# tunefs.lustre --dryrun lpool-oss/ost0 checking for
TR> existing Lustre data: found
TR> Read previous values:
TR> Target: testfs-OST0000 Index: 0 Lustre FS: testfs Mount type:
TR> zfs Flags: 0x62
TR> (OST first_time update )
TR> Persistent mount opts: Parameters: mgsnode=10.20.0.3@o2ib
TR> ost.quota_type=ug ...
TR> So indeed, the information about the first MGS NID
TR> (--mgsnode=10.20.0.2@o2ib) was lost somehow.
TR> Changing the failover NID specification to 'colon-syntax' does
TR> the trick: OSS:~# mkfs.lustre --reformat --ost --backfstype=zfs
TR> --fsname=testfs --index=0 --param ost.quota_type=ug
TR> --mgsnode=10.20.0.2@o2ib:10.20.0.3@o2ib lpool-oss/ost0 raidz2
TR> dm-1 dm-2 dm-3 ...
TR> Mounts smoothly.
TR> The example in the manual suggests that the doubling of
TR> '--mgsnode' would still work:
TR>
https://build.hpdd.intel.com/job/lustre-manual/lastSuccessfulBuild/artifa...
TR> A flaw in the manual, or did I miss something?
TR> Regards, Thomas
TR> _______________________________________________ HPDD-discuss
TR> mailing list HPDD-discuss(a)lists.01.org
TR> https://lists.01.org/mailman/listinfo/hpdd-discuss
--
Roland
-------
http://www.q-leap.com / http://qlustar.com
--- HPC / Storage / Cloud Linux Cluster OS ---