Online Backups Cross Platform?

[T.A.C.]

I'm trying to migrate an existing cluster from a Solaris 10 x86_64 environment to a RedHat Linux x86_64 environment (both running Opteron based servers).  The nodes in each environment are running 5.0.41; the Solaris cluster is running 3 nodes (which I know now is not recommended, but this was setup nearly 2 years  ago and has been running fine) and the Linux cluster is running 2 nodes.  My understanding is that you can still use a restore when changing the number of replicas.

I've tried doing a restore from the Solaris-based backups on the Linux-based cluster and get the following:

/usr/bin/ndb_restore -b 631 -m -n 3 /tmp/BACKUP-631
Backup Id = 631
Nodeid = 3
backup path = /tmp/BACKUP-631
Ndb version in backup files: Version 5.0.41
Unable to connect with connect string: nodeid=0,localhost:1186
Retrying every 5 seconds. Attempts left: 12 11 10 9 8 7

I shutdown all the API nodes before trying this and otherwise the new cluster is starting fine so I don't think it's a network connectivity issue:

ndb_mgm> show
Cluster Configuration
---------------------
[ndbd(NDB)]     2 node(s)
id=2    @10.100.115.21  (Version: 5.0.41, Nodegroup: 0, Master)
id=3    @10.100.115.22  (Version: 5.0.41, Nodegroup: 0)

[ndb_mgmd(MGM)] 1 node(s)
id=1    @10.100.115.20  (Version: 5.0.45)

[mysqld(API)]   2 node(s)
id=4 (not connected, accepting connect from any host)
id=5 (not connected, accepting connect from any host)

Any ideas?

[Geert Vanderkelen]

Hi Todd,

On 24 Jul 2007, at 11:48, T.A.C. wrote:
..

> /usr/bin/ndb_restore -b 631 -m -n 3 /tmp/BACKUP-631
> Backup Id = 631
> Nodeid = 3
> backup path = /tmp/BACKUP-631
> Ndb version in backup files: Version 5.0.41
> Unable to connect with connect string: nodeid=0,localhost:1186
> Retrying every 5 seconds. Attempts left: 12 11 10 9 8 7
>
> I shutdown all the API nodes before trying this and otherwise the new
> cluster is starting fine so I don't think it's a network  
> connectivity issue:

Well, it says it can't connect to localhost:1186. Is the management  
server running on the host you are trying to run ndb_restore? If not,  
specify the connect string.

But anyway, here is a bug report about endian compatibility and backups:
  http://bugs.mysql.com/bug.php?id=27543
which might be important for you.

Cheers,

Geert

--
Geert Vanderkelen, Senior Support Engineer
MySQL GmbH, Germany, www.mysql.com





--
MySQL Cluster Mailing List
For list archives: http://lists.mysql.com/cluster
To unsubscribe:    http://lists.mysql.com/cluster?unsub=lists@...

[T.A.C.]

Geert,

That was indeed the issue.

Thanks for the heads up on the endian bug; that hopefully won't be an issue since both are x86_64 little endian but we will definitely check on that.

On Tuesday 24 July 2007 08:57pm, Geert Vanderkelen wrote:

> Hi Todd,
>
> On 24 Jul 2007, at 11:48, T.A.C. wrote:
> ..
>
> > /usr/bin/ndb_restore -b 631 -m -n 3 /tmp/BACKUP-631
> > Backup Id = 631
> > Nodeid = 3
> > backup path = /tmp/BACKUP-631
> > Ndb version in backup files: Version 5.0.41
> > Unable to connect with connect string: nodeid=0,localhost:1186
> > Retrying every 5 seconds. Attempts left: 12 11 10 9 8 7
> >
> > I shutdown all the API nodes before trying this and otherwise the new
> > cluster is starting fine so I don't think it's a network
> > connectivity issue:
>
> Well, it says it can't connect to localhost:1186. Is the management
> server running on the host you are trying to run ndb_restore? If not,
> specify the connect string.
>
> But anyway, here is a bug report about endian compatibility and backups:
>   http://bugs.mysql.com/bug.php?id=27543
> which might be important for you.

He said that both the Solaris and the Linux systems were running on the same
processor platform.  Thus, endianess could not be an issue or even a
contributing factor.  If his Solaris systems were on SPARC or SPARC64, then
it would have made sense to include this info.

Be that as it may, I'm kinda glad to have the reference, just so that I know
more about it, too.  So, thanks :) .
--
Lamont Peterson <peregrine@...>
Founder [ http://blog.OpenBrainstem.net/peregrine/ ]
GPG Key fingerprint: 0E35 93C5 4249 49F0 EC7B  4DDD BE46 4732 6460 CCB5
  ___                   ____            _           _
 / _ \ _ __   ___ _ __ | __ ) _ __ __ _(_)_ __  ___| |_ ___ _ __ ___
| | | | '_ \ / _ \ '_ \|  _ \| '__/ _` | | '_ \/ __| __/ _ \ '_ ` _ \
| |_| | |_) |  __/ | | | |_) | | | (_| | | | | \__ \ ||  __/ | | | | |
 \___/| .__/ \___|_| |_|____/|_|  \__,_|_|_| |_|___/\__\___|_| |_| |_|
      |_|               Intelligent Open Source Software Engineering
                              [ http://www.OpenBrainstem.net/ ]