Restore (ontape -r) problems
Hi
We have two SPARC machines running Solaris 8 & IDS 7.31.UD6. One is a
Sun (test), the other a Futjitsu(live).
The test box has an instance of online running, with an exact copy of
the onconfig from the live box. All that has been changed is
commenting out the DBTEMP lines, and swapping the two tape devices over
as only /dev/rmt/1 is a DDS4 drive. However although all the logical
bits of disk are present only the rootdbs is initialised. We have not
created the other database spaces & chunks
There is an identical set of symbolic links to the actual bits of disks
on both boxes, and the kernel parameters have been altered in the same
way on both.
So, stopped IDS (having managed to initialise it) on the test box, put
a level 0 tape from the live box into the drive, and ran 'ontape -r'.
(logged in as 'informix') The archive information was displayed, then a
correct list of spaces to restore, then a correct list of dbspaces,
then a correct list of chunks. Then this:
===============================
Continue restore? (y/n)y
Do you want to back up the logs? (y/n)y
Please mount tape 1 on /dev/rmt/0n and press Return to continue ...
Would you like to back up log 1? (y/n) y
Please label this tape as number 1 in the log tape sequence.
This tape contains the following logical logs:
1
Log salvage is complete, continuing restore of archive.
Physical restore failed - function read archive backup failed code 1
errno 2
Program over.
================================
Frustration!
Is this a tape drive problem? (no sign of it in the messages file)
Or a tape problem? Clearly it has *started* to read the tape.
Or because although *both* machines are SPARC Solaris 8 IDS 7.31.UC6
(and the RDS was simply copied over & seems to work) they are *still*
not similar enough?
Any advice welcome...
Unless I am mistaken code 1errno 2 means 'file not found'. be aware
that ontape -rcan only restore chunks that are already present. This
means you MUST recreate the entire instance with all the chunks present
with the same sizes BEFORE you do the restore.
Let me know Iif I can help.
Regards,
Robert A. Reissaus
Risdi, Amsterdam, Barcelone, Gent, Zürich
IBM / Informix consultants since (ooo, way to long)
r.a.reissaus@risdi.com
www.risdi.com
Cats wrote:
> Hi
>
> We have two SPARC machines running Solaris 8 & IDS 7.31.UD6. One is a
> Sun (test), the other a Futjitsu(live).
>
> The test box has an instance of online running, with an exact copy of
> the onconfig from the live box. All that has been changed is
> commenting out the DBTEMP lines, and swapping the two tape devices over
> as only /dev/rmt/1 is a DDS4 drive. However although all the logical
> bits of disk are present only the rootdbs is initialised. We have not
> created the other database spaces & chunks
>
> There is an identical set of symbolic links to the actual bits of disks
> on both boxes, and the kernel parameters have been altered in the same
> way on both.
>
> So, stopped IDS (having managed to initialise it) on the test box, put
> a level 0 tape from the live box into the drive, and ran 'ontape -r'.
> (logged in as 'informix') The archive information was displayed, then a
> correct list of spaces to restore, then a correct list of dbspaces,
> then a correct list of chunks. Then this:
>
> ===============================
> Continue restore? (y/n)y
> Do you want to back up the logs? (y/n)y
> Please mount tape 1 on /dev/rmt/0n and press Return to continue ...
/dev/rmt/0n is a non-rewind device.
So, this is what you have for LTAPEDEV.
What have you got for TAPEDEV??
> Would you like to back up log 1? (y/n) y
> Please label this tape as number 1 in the log tape sequence.
> This tape contains the following logical logs:
> 1
> Log salvage is complete, continuing restore of archive.
> Physical restore failed - function read archive backup failed code 1
> errno 2
> Program over.
> ================================
>
> Frustration!
>
> Is this a tape drive problem? (no sign of it in the messages file)
>
> Or a tape problem? Clearly it has *started* to read the tape.
>
> Or because although *both* machines are SPARC Solaris 8 IDS 7.31.UC6
> (and the RDS was simply copied over & seems to work) they are *still*
> not similar enough?
>
> Any advice welcome...
>