RMAN-00571: ===========================================================
RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS
===============
RMAN-00571: ===========================================================
RMAN-03009: failure of backup command on edmptape channel at
11/04/2003 13:34:27
ORA-19506: failed to create sequential file, name="b_1sf5jlj1_1_1",
parms=""
ORA-27007: failed to open file
ORA-19511: Error received from media manager layer, error text:
SBT error = 7009, errno = 59018000, sbtopen: can't connect with
media manager
Oracle support says we have to look at the media layer. But there is
not much to look. We enabled tracing but basically the same error is
displayed in the agent trace.
Is there anyone out there that got this successfully to work and can
give us a hint what we doing wrong here.
The database and actually the whole server was restarted after the
installation of the arcserve agent.
Well, I know it's not what you are actually after, but why people backup
directly to tape from RMAN I've never figured out. It's *good* to have
backups on disk. Quicker to restore from, for one thing.
Any old tape backup utility can backup an RMAN disk-based backupset to tape.
Only the *expensive* (and sometimes flakey) stuff does it directly to tape
via the MML.
Besides, I never did get on with ArcServe.
Regards
HJR
Because I've got a multi-terabyte database and I just don't have the
disk space to hold the backup pieces. That's why I go directly to tape.
Tis better to have a backup that is slow to recover from than no backup
at all! But for smaller databases, I definitely go directly to disk and
then *copy* the backup to tape later.
Cheers,
Brian
--
===================================================================
Brian Peasland
dba@remove_spam.peasland.com
Remove the "remove_spam." from the email address to email me.
"I can give it to you cheap, quick, and good. Now pick two out of
the three"
You are totally right. And from experience, I can tell you that Latest
versions of ArcServe (now brightstor) is VERY flakey (At least on netware)
and support is awful and/or non-existent.
We are currently using an antedeluvian version of Cheyenne ArcServe/Open
(2.2c) that is no longer supported but is still working by miracle under
solaris 2.9. The big problems started to appear when CA took over cheyenne
software.
We might be moving to Veritas Net Backup (VERY expensive) but we will not
use the RMAN agent.
As long as your disk backups are on a seperate system I would agree.
Backing up to a SAN for example. However in our case we don't have
that here and my disk backups are ending up on the same system as my
database is sitting on. And in the case of a total loss of your disk
array (what happened twice to us with RAID5 arrays) you just wiped out
your backups too. Even if you backup first to disk and then with
whatever tool to tape - in case of a total disk array loss you just
lost your most recent archive redos too. This is why I want to back up
at least the archive redos on a frequent basis directly to tape in
addtion having them on disk.
Anyway my problem is still not solved: any other input?
HTH,
It's only disk space.
And as we all know (everyone can join in the chorus at this point)
"Diskspace is cheap".
And I've got my Seagate shares to worry about.
;-)
(Seriously: of course when you get to the multi-hundred Gigabyte sized
database, you have little choice. I like the robot arms that will load the
tapes for me, too!!).
Regards
HJR
"Brian Peasland" <dba@remove_spam.peasland.com> wrote in message
news:3FA92229.F1ECFA@remove_spam.peasland.com...
Regards
HJR
"Brian Peasland" <dba@remove_spam.peasland.com> wrote in message
news:3FA9332D.4A8B3B09@remove_spam.peasland.com...
> We are trying to use Arcserve 9 and its Oracle Backup agent with RMAN.
> We configured the RMAN oracle configuration in Arcserve but the rman
> job is constantly failing with this errorstack:
>
> RMAN-00571: ===========================================================
> RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS
> ===============
> RMAN-00571: ===========================================================
> RMAN-03009: failure of backup command on edmptape channel at
> 11/04/2003 13:34:27
> ORA-19506: failed to create sequential file, name="b_1sf5jlj1_1_1",
> parms=""
> ORA-27007: failed to open file
> ORA-19511: Error received from media manager layer, error text:
> SBT error = 7009, errno = 59018000, sbtopen: can't connect with
> media manager
It's quite a while ago but I had similar problems with tsm. Did you check the
logfile (in udump or bdump directory). In my old case it had to do with
incorrect location and contents of the tsm admin files ...
--
With kind regards,
Ronald
http://ronr.nl/unix-dba
It sure is. However...tried to purchase 10TB lately? It aint that
cheap!!!! LOL.
> (Seriously: of course when you get to the multi-hundred Gigabyte sized
> database, you have little choice. I like the robot arms that will load the
> tapes for me, too!!).
And that's what we did. Coupled RMAN with our existing, in-house Legato
managed, automated tape library silo. Just had to spend $6K for the
Legato Module that lets RMAN talk to Legato, and vice versa. Much
cheaper than buying disk!
Brian,
I hate to ask, but - how long is a full restore of rman level 0, plus
whatever incrementals and the archived logs - of your worst case
scenario?
Are you sure that NAS over gigabit wouldn't supply a lower mean time
to recovery?
How much is a lower MTTR worth, anyways?
Paul
I don't want to know the answer to that question, to be honest.
> Are you sure that NAS over gigabit wouldn't supply a lower mean time
> to recovery?
It probably would. However, management doesn't want to spend money for a
disk-based B&R solution, at this time. They know the problems with the
solution, such as the amount of time to fully recover. And they accept
the risk too. And just to make sure, I have emails for CYA type of
thing.
HTH,
hardcopy. :D
Those are exactly the kind of things that you don't want to have
"autoarchived" out of your inbox.
Pd