imsbackup > imsrestore via | ssh
Can someone that has mucked with an interactive store transfer using imsbackup >> imsrestore tell me if the error i am getting here is on the imsbackup end or imsrestore end , perhaps b/c of a syntax error with imsrestore ?
I am using the syntax right out of the admin guide, and I have confirmed that my imsbackup syntax works before piping it through an ssh shell.
What am I missing ?
bash-2.03# ./imsbackup -f - -u test.group | ssh -l john milonga.education.ucsb.edu /opt/SUNWmsgsr/lib/msg/imsrestore.sh -f- /primary -cy -v1
[22/Jan/2007:14:16:28 -0800] milonga /opt/SUNWmsgsr/sbin/imsrestore[23377]: General Critical: Unable to verify start of store daemon (stored). The store daemon process must be running and the initialization must be finished before this process starts. If the store daemon is running, do not restart it; wait until it has finished initializing, then start this process again. If the store daemon is not running, please start it.
Unable to verify start of store daemon (stored). The store daemon process must be running and the initialization must be finished before this process starts. If the store daemon is running, do not restart it; wait until it has finished initializing, then start this process again. If the store daemon is not running, please start it.
unable to backup: /minerva/ALL/john/INBOX
Device error
Error occured, see default logfile
[1440 byte] By [
goubeauxa] at [2007-11-26 15:52:42]
# 1
Hi,
I have to ask the obvious question, given the error message, were the messaging server instances running on both systems when you ran the test?
If they were and imsrestore couldn't 'detect' they were running, this may be due to incorrect path/environment settings.
Also I notice that you login as user 'john' to milonga, does this user have enough permissions, shouldn't you be using something like mailsrv instead?
Another hint to reduce cpu overhead of using ssh is to use ssh -c blowfish. This uses the blowfish algorithm, which is less secure but a lot faster/less cpu intensive.
Regards,
Shane.
# 2
> I have to ask the obvious question, given the error
> message, were the messaging server instances running
> on both systems when you ran the test?
Yes, indeed, in fact given the error, i stopped and started the stored on the JES instance just in case it was in a odd state.
>
> If they were and imsrestore couldn't 'detect' they
> were running, this may be due to incorrect
> path/environment settings.
Ok, this might be b/c of the user i am trying this as, as well?
>
> Also I notice that you login as user 'john' to
> milonga, does this user have enough permissions,
> shouldn't you be using something like mailsrv
> instead?
This is likely the problem, though i added the user <john> to the group that should be able to run imsrestore. I will try it as mailsrv to be sure.
>
> Another hint to reduce cpu overhead of using ssh is
> to use ssh -c blowfish. This uses the blowfish
> algorithm, which is less secure but a lot faster/less
> cpu intensive.
Good tip, will do this when i move a large number of data
>
> Regards,
>
> Shane.
# 3
>
>
> Ok, this might be b/c of the user i am trying this
> as, as well?
> >
> > Also I notice that you login as user 'john' to
> > milonga, does this user have enough permissions,
> > shouldn't you be using something like mailsrv
> > instead?
>
> This is likely the problem, though i added the user
> <john> to the group that should be able to run
> imsrestore. I will try it as mailsrv to be sure.
>
Ok,
it was the fact that the user did not have proper perms. and after changing to the mailsrv user it works fine.
Thanks for the insights
-john
> >
> > Another hint to reduce cpu overhead of using ssh
> is
> > to use ssh -c blowfish. This uses the blowfish
> > algorithm, which is less secure but a lot
> faster/less
> > cpu intensive.
>
> Good tip, will do this when i move a large number of
> data
>
> >
> > Regards,
> >
> > Shane.
