Networker

Re: [Networker] W2K Client Problems savepnpc tmp file exists

2003-01-21 06:52:25
Subject: Re: [Networker] W2K Client Problems savepnpc tmp file exists
From: Davina Treiber <treiber AT HOTPOP DOT COM>
To: NETWORKER AT LISTMAIL.TEMPLE DOT EDU
Date: Tue, 21 Jan 2003 06:52:29 -0500
On Tue, 21 Jan 2003 11:03:42 +0100, Andre Beck <networker AT IBH DOT DE> wrote:

>A bit late, but:
>
>On Mon, Nov 04, 2002 at 08:46:11AM -0500, Robert Maiello wrote:
>> On Unix 6.1.2  server and client savepnpnc will ALWAYS produce the
>> output you have here.   Pre and Post processing run fine however.
>
>Not exactly. The difference beeing here is that with 6.2, the failing
>of subsequent savesets but the first of a savepnpc group - this means
>the "failure" of preclntsave to write the (existing) tmp file is inter-
>preted as fatal and savesets are *not* saved.

As I said in my earlier posting, you're probably best not to use 6.2 unless
you really need to. I haven't tried 6.2 but it's interesting to know that
they have broken savepnpc again.

>
>> Legato told me this not a bug in 6.1.2 but a feature and that it is
>> working as designed ??  (No fix or workaround will be provided).
Not quite working as designed. This was a failed attempt to fix a long
running bug, first reported as LGTpa06002. They have had several attempts at
this and failed dismally each time. These messages in 6.1.2 are totally
pointless, all they do is identify the existence of the tmp file. On the
first run through, the tmp file should not exist, and on subsequent save
sets it is meant to be there. How do they think a message like this is going
to help anyone know whether it is correct? The bug ID here was LGTpa35288,
and the 6.1.2 fixed bugs list says "fixed". WRONG. The same bug ID is also
listed in the 6.1.3 fixed bugs list as "fixed". WRONG again - I have tested
it and it is still broken, they just changed the wording of the message.

My guess is that on several occasions they have given the task of fixing
this bug to a junior programmer who really doesn't understand how savepnpc
works. Recently I got really tired of this appalling lack of progress and
contacted someone I know in Legato Engineering who is extremely
knowledgeable and experienced, and I asked him to fix it. To my delight he
took it on board, and came up with a proper solution to the problem. He
raised a bug (LGTpa49608) and this is expected to be fixed in 7.x. I have
high hopes that it WILL work this time.

In the mean time, the best workaround is to use savefixpnpc (see the
NetWorker FAQ-o-matic). This is currently available for Unix, and I have
been working on the NT version. I believe this is now working but I am still
in the final testing phase. I could release on a beta basis if anyone is
desperate for it.

>
>The somewhat superfluous messages will probably not get fixed, as it is
>hard to differ whether they are actually superfluous or it would be better
>to warn the admin - on saveset level, where the preclntsave runs, this
>is not really easy to decide.
They are superflous, they don't serve any useful purpose except to confuse
newbies and experienced users alike.

>
>The error, however, is that the saveset will fail:
>
>
>> >* sec-ws-03:C:\ 1 retry attempted
>> >* sec-ws-03:C:\ preclntsave: The tmp file exists. Another savepnpc may be
>> running.
>
>This is normal.
>
>> >* sec-ws-03:C:\  Preclntsave exiting......!
>
>But this is not.
Isn't it? I haven't tried 6.2 myself, but my guess would be that preclntsave
exiting is not a problem, unless it exits with a non-zero exit code. If it
does this (I suppose that is what you are saying) then Legato have excelled
once again in the incompetence stakes.

>
>I've opened a case on that problem and got a fixed version instantly the
>next day. It is a modified preclntsave.exe which, when replaced for the
>original on the savepnpc client, lets savepnpc work as expected. You
>can find it (for NT/2k/XP, but with 6.2 that's the only relevant platform
>anyway) at ftp://ftp.legato.com/outgoing/3041931/ (or you may use the
>case number when refering to your own problem. I was told 6.2.1 will have
>the fix, too.
Hmm, methinks 6.2 is well worth avoiding.

--
Note: To sign off this list, send a "signoff networker" command via email
to listserv AT listmail.temple DOT edu or visit the list's Web site at
http://listmail.temple.edu/archives/networker.html where you can
also view and post messages to the list.
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=