I finally figured out what was the problem. It turned out that Solaris 8
does not have a current patch of ufsrestore which correct the "xtrmap: too
many map entries" error. That error occurs only if your filesystem has so
many files in it. Because of that error, index entries were not
piped through for amanda index server. There are patches for Solaris
2.6 and Solaris 7 on sunsolve.sun.com but there are no patches available
for Solaris 8. However, the patches for Solaris 8 is available ONLY to
those who have support contract with Sun. While I wait for the patches
from Sun, I copied the patched ufsrestore from Solaris 2.6 and it worked
well so far.
Josh
"John R. Jackson" wrote:
> >Here is what it looks like in my disklist. ...
> >and in index directory of homer:
>
> Thanks.
>
> So the index area is being created properly and the files are being
> created, they are just empty (which is what you said, but I wasn't clear
> about the details).
>
> It might be that the index generation on the client is failing and
> not sending anything across the pipe, or the server side may have
> done something wrong, or the network connection may have been lost
> (or something else :-).
>
> I'd look in two places for more information. First is amdump.<nn> on the
> tape server. Look for any out of the ordinary messages during the time
> this file system was being backed up (between the FILE-WRITE/PORT-WRITE
> and the time the dumper completes).
>
> The other place I'd look is in the /tmp/amanda/sendbackup*debug file
> on the client. But you'll have to catch one with the dump of this file
> system, which may be tricky. I build Amanda with --with-pid-debug-files
> so I can keep them apart. Look for any error messages and also that
> the right restore program is being used and with the right arguments.
>
> John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
