Re: [Networker] Solaris to Linux Migration
2007-08-14 10:27:12
Please corroborate this with others on this listing - I hope they jump
in - but it's my understanding that it SHOULD be done from tape. The
reason, or rather one possible reason, is because you are going from one
type of endianess to another. Linux is not the same endian as Solaris.
One is big, the other is little. Can't recall which is which, and it
doesn't really matter, but that *could* be a potential gotcha. It's
never been clear to me whether the Legato software is written to care
about that or take advantage of that, and I've never been able to get a
confirmed (averred) answer one way or the other from Legato. It all
depends on how the software interprets it. Some software is written so
as not to care, other packages would care or do use the endianess to its
advantage.
It's also my understanding that if you back up and recover from tape
then the OS is not an issue and is essentially removed from the equation
as the format on tape is neutral to this. As I recall, the client
indexes can be manually copied using something like rsync, or whatever,
but the database is OS dependent and must be backed up and recovered
from tape to obviate this problem. I don't think recovering 70 GB should
take that long.
The main thing here, though, is that EMC-Legato does not support
migrating from Solaris to Linux. The last time I checked, they did not
officially sanction this. However, they say that many have done it, and
that if you carry out the procedure as you would when moving from any
server to a new server running that same OS, ie. using the standard
methodology, then it *SHOULD* work. Others who have done this claim it
works, but I can't imagine they've done extensive testing, so maybe a
few spot recover tests here and there and a few queries against the
media database to make sure you get the same results. Extensive testing
would take more time than most of us have, however. For example, if
you're upgrading at the same time then you will undoubtedly have a newer
version of mminfo wherein it's entirely possible that certain queries
could produce different results on the new server, particularly when
running in verbose mode or something that will generate more detail.
If I were doing this, I would probably manually copy the indexes and
then backup/recover the database to the new Linux server, run some
queries on the new Linux server against the media database and compare
the results with what I generated on the old Solaris box and then also
run a few test recovers, using both nwrecover and save set recover,
maybe clone some older data and a few new backups and recovers and leave
it at that.
I think you just do it, test a few things, and if all looks well, you
just hope no older legacy data becomes inaccessible later or breaks
further down the road. Another way to look at it is that you will never
be 100% certain. No way. But, you will feel better as time passes,
knowing that most recovers usually are for more recent data, which in
this case would be data that was backed up on the new host. Little
succor perhaps, but what can you do? Sooner or later, you have to set
sail and leave the safe harbor behind.
George
Markus Wells wrote:
Hi all,
I'm new here on the Networker list. I scanned the archives and couldn't
find much info for the issue I have at hand so I decided to post it directly.
The issue is that when we tried to migrate networker from Solaris to Linux,
it failed due to being unable to acccess the built-in volume/label/whatever
database. the person who actually did it believes it had something to do
with the database going from a UFS to an ext3 filesystem and Legat just not
liking that.
I'm going to try it again, but I wanted to ask if anyone has ever
encountered a problem like that and what a possible solution might be.
I contacted Matt Hemingway who posted about a similar migration on here and
he said he had no issues and thought it might be due to his restoring the
database from tape rather than doing a rsync or scp.
One other thing I should mention is that our "database" (or whatever it's
called in Legato/EMC speak) is over 70G as we have some 2-3000 tapes cataloged.
Any thought, hints, RTFM pointers, etc greatly appreciated.
Thank you,
Markus
To sign off this list, send email to listserv AT listserv.temple DOT edu and type
"signoff networker" in the body of the email. Please write to networker-request
AT listserv.temple DOT edu if you have any problems with this list. You can access the
archives at http://listserv.temple.edu/archives/networker.html or
via RSS at http://listserv.temple.edu/cgi-bin/wa?RSS&L=NETWORKER
--
George Sinclair - NOAA/NESDIS/National Oceanographic Data Center
SSMC3 4th Floor Rm 4145 | Voice: (301) 713-3284 x210
1315 East West Highway | Fax: (301) 713-3301
Silver Spring, MD 20910-3282 | Web Site: http://www.nodc.noaa.gov/
- Any opinions expressed in this message are NOT those of the US Govt. -
To sign off this list, send email to listserv AT listserv.temple DOT edu and type
"signoff networker" in the body of the email. Please write to networker-request
AT listserv.temple DOT edu if you have any problems with this list. You can access the
archives at http://listserv.temple.edu/archives/networker.html or
via RSS at http://listserv.temple.edu/cgi-bin/wa?RSS&L=NETWORKER
|
|
|