That does help. I am in a SAN environment with SSO. However, this
particular box is just a client. My other SAN media servers are Solaris
2.8 and I am utilizing ALTREADHOST and doing all the duplication from my
master.
Thanks,
Brad Hillebrand
FHLBI
-----Original Message-----
From: david AT dsihost-srv01 DOT com [mailto:david AT dsihost-srv01 DOT com]
Sent: Tuesday, June 11, 2002 2:41 PM
To: Hillebrand, Bradly W.
Subject: RE: [Veritas-bu] Vault errors
I'm assuming the remote host here is a media server (the 2000 box), so
yes,
that's true, but there is a socket created for the bptm processes (one
for read
and one for write) to communicate with each other.
Not knowing how you have your environment set up (SSO, etc.), this is
how
bpduplicate ( taking vault out of the discussion for right now )
operates.
When bpduplicate starts up a process to duplicate an image, it will,
based on
the image or media, determine the media server that originally wrote
this image
to tape. At that point, it makes a request to that media server to
mount that
tape. Once that is mounted, it then makes a call to bptm to read the
image
data from that tape and send it back to the requesting server (in this
case the
master). On the master's end bpduplicate also made a request for a tape
to be
mounted and another bptm process responsible for writing the received
image is
spawned. This communication continutes until the image has been
successfully
duplicated. If at any point during this process, the network connection
is
reset, or cut for some reason, bpduplicate or rather bptm, will complain
that
it lost communication with remote host. Therefore the duplication
fails.
If you do have an SSO environment, where you are sharing your tape
drives, then
you should be able to implement the altreadhost option in bpduplicate.
This
bascially says, even if the tape is owned by another media server, if
the
master can mount and read the tape let it be the alternate reading host
to
minimize network (LAN) traffic).
Have you seen many 13/14 errors with regard to this 2000 media server
during
regular backups?
Let me know if this helps.
David Chapa
Quoting "Hillebrand, Bradly W." <bhillebrand AT fhlbi DOT com>:
> The remote host is a 2000 box, but I don't really understand that as
> well. What would a vault be doing on any remote host? Is it not just
> reading images from tape#1 and writing them to tape#2?
>
> Thanks,
>
> Brad Hillebrand
> FHLBI
>
>
> -----Original Message-----
> From: david AT dsihost-srv01 DOT com [mailto:david AT dsihost-srv01 DOT com]
> Sent: Tuesday, June 11, 2002 11:11 AM
> To: Hillebrand, Bradly W.
> Cc: veritas-bu AT mailman.eng.auburn DOT edu
> Subject: Re: [Veritas-bu] Vault errors
>
> Did you happen to have any logging turned on (ie. bpdbm) you can look
> at?
>
> Can you replicate this problem?
>
> I would wonder what was happening on the remote host that would have
> caued this
> kind of error.
>
> Is it an NT server?
>
> Could be a Networking issue as well.
>
> sorry for the rapid fire, but there are so many things it MIGHT be.
>
> David Chapa
> DataStaff, Inc.
>
> Quoting "Hillebrand, Bradly W." <bhillebrand AT fhlbi DOT com>:
>
> >
> >
> > I am getting the following errors in my bpvault sessions:
> >
> >
> >
> > bpduplicate[1]: ginkgo1[1]: 01:49:32 INF - socket read failed: errno
=
> > 10054 - An existing connection was forcibly closed by the remote
host.
>
> >
> >
> >
> >
> >
> > After this happens I get the following message(s) until I kill the
> jobs:
> >
> >
> >
> > bpduplicate[0]: ginkgo1[0]: DUP33: 06/10/02 02:19:33 no duplicate
> > progress message in 30 minutes
> >
> >
> >
> > Any ideas?
> >
> >
> >
> > Thanks,
> >
> >
> >
> > Brad Hillebrand
> >
> > FHLBI
> >
> >
> >
> >
>
>
>
|
