Rman restore : failed to retrieve

Please attach script to check, and a simple topology. Aditional check in /etc/hosts (or C:\windows\system32\drivers\etc\hosts) if exist networker server, host and host where runing backup

hosts are well informed on 2 sides.

i'm trying to restore on client2 through networker2 a database named BASE from client1 from networker1.
scanner is good.
nsrinfo well raised informations about client1 on networker2, mediadatabase also.

networker2:>nsrinfo -n oracle client1
scanning client `client1' for all savetimes from the oracle namespace
ORA_CFc-609576117-20090115-0d, date=1232036727 Thu 15 Jan 2009 05:25:27 PM CET
/Arch_t676229120_s219_p1/, date=1232036722 Thu 15 Jan 2009 05:25:22 PM CET
ORA_CFc-609576117-20090115-0c, date=1232036710 Thu 15 Jan 2009 05:25:10 PM CET
/FULL_t676229104_s217_p1/, date=1232036705 Thu 15 Jan 2009 05:25:05 PM CET
/FULL_t676228828_s216_p1/, date=1232036430 Thu 15 Jan 2009 05:20:30 PM CET
5 objects found


rman script used on client2 :
connect target /
run {
allocate channel B1 type 'SBT_TAPE';
send channel B1 'NSR_ENV=(NSR_CLIENT=client1,NSR_DATA_VOLUME_POOL=MIGTST,NSR_SERVER=networker2)';
# parms 'ENV=(NSR_CLIENT=csga0003)';
restore database;
release channel B1;}

i'm using information from control files and not rman catalog, but doesn't matter i presume.

result :
connected to target database: BASE (DBID=609576117, not open)

using target database control file instead of recovery catalog
allocated channel: B1
channel B1: sid=40 devtype=SBT_TAPE
channel B1: NMO v4.5.0.0

sent command to channel: B1

Starting restore at 22-JAN-09

channel B1: starting datafile backupset restore
channel B1: specifying datafile(s) to restore from backup set
restoring datafile 00001 to /oracle/BASE/data1/system01.dbf
restoring datafile 00002 to /oracle/BASE/data5/undotbs01.dbf
restoring datafile 00003 to /oracle/BASE/data2/sysaux01.dbf
restoring datafile 00004 to /oracle/BASE/data2/tools01.dbf
restoring datafile 00005 to /oracle/BASE/data1/patrol.dbf
restoring datafile 00006 to /oracle/BASE/data3/tab_sgcnr_01.dbf
restoring datafile 00007 to /oracle/BASE/data4/idx_sgcnr_01.dbf
channel B1: reading from backup piece /FULL_t676228828_s216_p1/
ORA-19870: error reading backup piece /FULL_t676228828_s216_p1/
ORA-19507: failed to retrieve sequential file, handle="/FULL_t676228828_s216_p1/", parms=""
ORA-27029: skgfrtrv: sbtrestore returned error
ORA-19511: Error received from media manager layer, error text:
Could not locate the LNM save file '/FULL_t676228828_s216_p1/' on server 'ces3rapbkp'. (2:9:2)
failover to previous backup

creating datafile fno=6 name=/oracle/BASE/data3/tab_sgcnr_01.dbf
released channel: B1
RMAN-00571: ===========================================================
RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS ===============
RMAN-00571: ===========================================================
RMAN-03002: failure of restore command at 01/22/2009 10:56:56
ORA-01119: error in creating database file '/oracle/BASE/data3/tab_sgcnr_01.dbf'
ORA-27038: created file already exists
Additional information: 1

Recovery Manager complete.

Hope it's clear.
thanks for your help

I see two errors... one about read and one about write. I guess second one is caused by your previous attempt which failed. I believe the easiest way to do this is to use RMAN copy command which would restore original database on target and that would actually what Oracle ppl would suggest. Can you try that?

you suggest to modify the rman backup script to do that.
So we don't use rman backup as usual...
The aim was to be able to recover a rman backup from different site.
We can't envisage to change all rman backup scripts.

damage.

thanks

We move tape from netw1 to netw2 and by a scanner on netw2 of tape from client1.
My client, pool is created by anticipation in netw2.
This well works as you can see the nsrinfo answer.

pardon your question were more about how rman know where find the backup.

I catch a controlfile from client1 near the backup occurance. So I ensure that i have the rman backup information in this ctlfile.

i mount a database with this ctl, replicating them according init.ora.... that works also. This database is started on client2 which is backuped by netw2.

So i run the script and that was good one time but now no way. And the message let thinking about a networker problem but perhaps inderctly a database pb ?

Actually the modification would be according to what Oracle would say and their practice.

But wait, I just noticed something else. You seem to talk about 2 backup server and 2 clients. Can you tell me how nw2 server knows about saveset from nw1 server?

I did this only using copy once. I assume mminfo for these saveset on both server is the same?

yes exactly.

Only scanner for tape? In networker_2 exist any index for old client?

Well, scanner -i would bring the index back, right? Besides, we know it is there due to nsrinfo output...

Both clients are created in new Networker Server?

client1 from netw1 (a datazone) has been created in netw2 (another datazone).
The recover occured from the netw2 on client2 (a client of netw2) but with data from client1, well known by the scanner.

I have the same problem a year ago. Check in client2 /etc/hosts exist any reference for client1. If not, add a new line with ip and name for client1.

don't be disapointed by different alias used for this post (tracks or jvb) we have 2 accounts, personnal (tracks) and corporate (jvb)....
thanks again !!