Sybase NNTP forums - End Of Life (EOL)

The NNTP forums from Sybase - forums.sybase.com - are now closed.

All new questions should be directed to the appropriate forum at the SAP Community Network (SCN).

Individual products have links to the respective forums on SCN, or you can go to SCN and search for your product in the search box (upper right corner) to find your specific developer center.

Runaway BCP

5 posts in Windows NT Last posting was on 1998-02-18 01:13:13.0Z
Vince Posted on 1998-02-13 16:56:03.0Z
From: "Vince" <verceg@intekinfo.com>
Subject: Runaway BCP
Date: Fri, 13 Feb 1998 08:56:03 -0800
Lines: 48
X-Newsreader: Microsoft Outlook Express 4.72.2106.4
X-MimeOLE: Produced By Microsoft MimeOLE V4.72.2106.4
Message-ID: <IirkCdJO9GA.128@forums.powersoft.com>
Newsgroups: sybase.public.sqlserver.nt
Path: forums-1-dub!forums-master.sybase.com!forums.powersoft.com
Xref: forums-1-dub sybase.public.sqlserver.nt:5194
Article PK: 1081273

I have a batch file that runs a series of BCP and ISQL commands in a Win95
dos-box. It has been doing its job for a couple of months with little
incident. Twice in the past two weeks, one of the BCP "out" commands has
hiccupped midway through, created a file over 2GB in size (the data I'm
requesting is *maybe* 20 megs), and creates an output file (I'm piping the
output to a text file) similar to this:

Starting copy...
1000 rows successfully bulk-copied to host-file.
2000 rows successfully bulk-copied to host-file.
3000 rows successfully bulk-copied to host-file.
4000 rows successfully bulk-copied to host-file.
5000 rows successfully bulk-copied to host-file.
6000 rows successfully bulk-copied to host-file.
7000 rows successfully bulk-copied to host-file.
CTLIB Message: - L5/O3/S5/N6/5/0:
ct_fetch(): network packet layer: internal net library error: Net-Library
operation terminated due to disconnect
CTLIB Message: - L5/O3/S5/N6/5/0:
ct_fetch(): network packet layer: internal net library error: Net-Library
operation terminated due to disconnect
CTLIB Message: - L5/O3/S5/N6/5/0:
ct_fetch(): network packet layer: internal net library error: Net-Library
operation terminated due to disconnect
CTLIB Message: - L5/O3/S5/N6/5/0:

(This continues on and on and on until this file reaches ~250MB and the bcp
data file is up over 2GB. The drive fills and I become very unpopular).

The server stats (SELECT @@version) are:
SQL Server/11.0.2.1/P/PC Intel/Windows NT 3.5/EBF 6722/OPT/Thu Oct 24
1996 19:59:07.06

This has happened in two separate tables, but from the same database. DBCC
shows no corruption. In both cases, I am bcp'ing out from a view instead of
a base table. And as I said before, this has been working for a couple of
months except for the past two weeks where this has happened twice.

BCP.EXE is dated 7-24-97 3:27pm 59,392 bytes.

I'm going to request my DBA open an official issue with Sybase, but it would
be great to hear if someone else has ever experienced this.

Vince Erceg
Intek Information Inc.
San Diego, CA


AnaMaria Stirbet Posted on 1998-02-16 22:41:32.0Z
Message-ID: <34E8C09C.5ED6@sybase.com>
Date: Mon, 16 Feb 1998 17:41:32 -0500
From: AnaMaria Stirbet <AnaMaria.Stirbet@sybase.com>
Reply-To: AnaMaria.Stirbet@sybase.com
X-Mailer: Mozilla 3.0 (WinNT; I)
MIME-Version: 1.0
Subject: Re: Runaway BCP
References: <IirkCdJO9GA.128@forums.powersoft.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Newsgroups: sybase.public.sqlserver.nt
Lines: 65
Path: forums-1-dub!forums-master.sybase.com!forums.powersoft.com
Xref: forums-1-dub sybase.public.sqlserver.nt:5186
Article PK: 1081265


Vince wrote:
>
> I have a batch file that runs a series of BCP and ISQL commands in a Win95
> dos-box. It has been doing its job for a couple of months with little
> incident. Twice in the past two weeks, one of the BCP "out" commands has
> hiccupped midway through, created a file over 2GB in size (the data I'm
> requesting is *maybe* 20 megs), and creates an output file (I'm piping the
> output to a text file) similar to this:
>
> Starting copy...
> 1000 rows successfully bulk-copied to host-file.
> 2000 rows successfully bulk-copied to host-file.
> 3000 rows successfully bulk-copied to host-file.
> 4000 rows successfully bulk-copied to host-file.
> 5000 rows successfully bulk-copied to host-file.
> 6000 rows successfully bulk-copied to host-file.
> 7000 rows successfully bulk-copied to host-file.
> CTLIB Message: - L5/O3/S5/N6/5/0:
> ct_fetch(): network packet layer: internal net library error: Net-Library
> operation terminated due to disconnect
> CTLIB Message: - L5/O3/S5/N6/5/0:
> ct_fetch(): network packet layer: internal net library error: Net-Library
> operation terminated due to disconnect
> CTLIB Message: - L5/O3/S5/N6/5/0:
> ct_fetch(): network packet layer: internal net library error: Net-Library
> operation terminated due to disconnect
> CTLIB Message: - L5/O3/S5/N6/5/0:
>
> (This continues on and on and on until this file reaches ~250MB and the bcp
> data file is up over 2GB. The drive fills and I become very unpopular).
>
> The server stats (SELECT @@version) are:
> SQL Server/11.0.2.1/P/PC Intel/Windows NT 3.5/EBF 6722/OPT/Thu Oct 24
> 1996 19:59:07.06
>
> This has happened in two separate tables, but from the same database. DBCC
> shows no corruption. In both cases, I am bcp'ing out from a view instead of
> a base table. And as I said before, this has been working for a couple of
> months except for the past two weeks where this has happened twice.
>
> BCP.EXE is dated 7-24-97 3:27pm 59,392 bytes.
>
> I'm going to request my DBA open an official issue with Sybase, but it would
> be great to hear if someone else has ever experienced this.
>
> Vince Erceg
> Intek Information Inc.
> San Diego, CA

Vince,
You are definitely running some 11.1.x OC because you have CT_lib
messages.
Now the Open Client on the Win95 machine was probably upgraded from
10.0.4 to 11.1 or 11.1.1
I'll suggest that you check if the upgrade has happend and compare the
behaviour of the two OC versions.
BTW: Are you using the -E switch or large numeric values?

Ana Stirbet


Mark A. Parsons Posted on 1998-02-14 00:46:00.0Z
Message-ID: <34E4E948.60B3@compuserve.com>
Date: Fri, 13 Feb 1998 19:46:00 -0500
From: "Mark A. Parsons" <Iron_Horse@compuserve.com>
Reply-To: Iron_Horse@compuserve.com
Organization: Iron Horse, Inc.
X-Mailer: Mozilla 3.01 (Win95; I)
MIME-Version: 1.0
Subject: Re: Runaway BCP
References: <IirkCdJO9GA.128@forums.powersoft.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Newsgroups: sybase.public.sqlserver.nt
Lines: 59
Path: forums-1-dub!forums-master.sybase.com!forums.powersoft.com
Xref: forums-1-dub sybase.public.sqlserver.nt:5191
Article PK: 1081271


Vince wrote:
> This has happened in two separate tables, but from the same database. DBCC
> shows no corruption. In both cases, I am bcp'ing out from a view instead of
> a base table. And as I said before, this has been working for a couple of
> months except for the past two weeks where this has happened twice.
>
> BCP.EXE is dated 7-24-97 3:27pm 59,392 bytes.
>
> I'm going to request my DBA open an official issue with Sybase, but it would
> be great to hear if someone else has ever experienced this.
>
> Vince Erceg
> Intek Information Inc.
> San Diego, CA

I thought you said you were getting rid of that pesky view on
vw_report_data!?!?!?! <g>

I've used bcp on views on 11.0.2.x dataservers with no problems so I
can't say I've ever had any problems with this scenario.

Curious ... the first CTLIB message mentions a 'disconnect'.

Any problems with networks lately ... specifically the one between your
Win95 PC and the dataserver maching? Can your network folks point to
any 'known' network delays during the times you had the problems?

I'm assuming this is intermittent and not reproducable on demand?

Is there anything different about the tables on which the bcp succeeds
and the tables on which the bcp fails, e.g., the tables that fail take
the longest to bcp from (lots rows and volume) while the ones that
succeed are relatively small in comparison? I'm wondering if any
lengthy (bcp) operation is hitting up against a timeout situation on
your PC or dataserver?

When you get these CTLIB messages ... can you tie it to a similar
message in the dataserver errorlog concerning a 'disconnected' spid? If
so, are there other occurrences of this same message in the dataserver
errorlog? If there are other occurrences of the same message in the
errorlog ... especially in the same time frame ... and you can trace
these disconnects back to specific users/PC/processes ... this may point
to a problem with a specific network resource?? (At my current client we
currently have one of router/hubs hiccup ... when that happens we'll see
a couple dozen spid's disconnected at the same time within the
dataserver errorlog.)

If none of these 'ideas' pan out ... sounds like a Microsoft OS bug!!
:-)
--

Mark Parsons
Iron Horse, Inc.
[Team Sybase]


Vince Erceg Posted on 1998-02-16 16:57:52.0Z
From: "Vince Erceg" <verceg@intekinfo.com>
References: <IirkCdJO9GA.128@forums.powersoft.com> <34E4E948.60B3@compuserve.com>
Subject: Re: Runaway BCP
Date: Mon, 16 Feb 1998 08:57:52 -0800
Lines: 31
Organization: Intek Information Inc.
X-Newsreader: Microsoft Outlook Express 4.72.2106.4
X-MimeOLE: Produced By Microsoft MimeOLE V4.72.2106.4
Message-ID: <eVdf9LvO9GA.421@forums.powersoft.com>
Newsgroups: sybase.public.sqlserver.nt
Path: forums-1-dub!forums-master.sybase.com!forums.powersoft.com
Xref: forums-1-dub sybase.public.sqlserver.nt:5188
Article PK: 1081269


>I thought you said you were getting rid of that pesky view on
>vw_report_data!?!?!?! <g>

Mark,

Not only did I get rid of the view, I got rid of the *data* that the view
was designed to filter out! Get rid of the offending data... no need for
the view. I think I've stumbled on a pretty revolutionary idea -- wonder if
it can be patented? :)

This is the same project, but a different server. The network guys have
checked both the NT and SQL error logs and nothing at all shows up at the
time that this happens. The problem is intermittent and not reproducable,
but thanks to the audit report, I can tell around what time this is
happening.

I set up a new machine with a fresh installation of the Sybase client
software to rule out a "bad" CTLIB dll or something. I'll see how that goes
tonight. In the interest of cutting my losses and being productive here,
I'm going to give it a couple more days and then I'm going to work on a VB
solution. At least I'll have a better chance at catching these types of
errors/issues than I do with a batch file.

Thanks for the ideas -- I'm not going to give up on this.

Vince Erceg
Intek Information Inc.
San Diego, CA


Mark A. Parsons Posted on 1998-02-18 01:13:13.0Z
Message-ID: <34EA35A9.317F@compuserve.com>
Date: Tue, 17 Feb 1998 20:13:13 -0500
From: "Mark A. Parsons" <Iron_Horse@compuserve.com>
Reply-To: Iron_Horse@compuserve.com
Organization: Iron Horse, Inc.
X-Mailer: Mozilla 3.01 (Win95; I)
MIME-Version: 1.0
Subject: Re: Runaway BCP
References: <IirkCdJO9GA.128@forums.powersoft.com> <34E4E948.60B3@compuserve.com> <eVdf9LvO9GA.421@forums.powersoft.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Newsgroups: sybase.public.sqlserver.nt
Lines: 22
Path: forums-1-dub!forums-master.sybase.com!forums.powersoft.com
Xref: forums-1-dub sybase.public.sqlserver.nt:5181
Article PK: 1081261


Vince Erceg wrote:
>

> Not only did I get rid of the view, I got rid of the *data* that the view
> was designed to filter out! Get rid of the offending data... no need for
> the view. I think I've stumbled on a pretty revolutionary idea -- wonder if
> it can be patented? :)

Sorry, dude ... I've already got a patent in on that idea. Actually, my
patent is much 'bigger' in that it goes further in suggesting that you
get rid of all databases and devices as well as all users (not just
their logins ... if ya know what I mean)!! I have seen incredibly
massive improvements in query processing times, minimal problems with
client code, extremely small foot prints for my dataservers ... and I
can run dbcc's to my heart's content with no complaints from the
'users'!!


--

Mark Parsons
Iron Horse, Inc.
[Team Sybase]