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.

Memory loss - Can someone explain this?

7 posts in General Discussion (old) Last posting was on 2000-02-27 07:36:55.0Z
Adam Simmonds Posted on 2000-02-26 00:42:48.0Z
Newsgroups: sybase.public.easerver
Date: Sat, 26 Feb 2000 11:42:48 +1100
From: Adam Simmonds <asimmond@mail.usyd.edu.au>
Organization: University of Syndey
X-Mailer: Mozilla 4.7 [en] (WinNT; I)
X-Accept-Language: en
MIME-Version: 1.0
Subject: Memory loss - Can someone explain this?
Content-Type: multipart/mixed; boundary="------------1877F6F402BFF83AA74E5625"
Lines: 44
NNTP-Posting-Host: pc-70.sis.usyd.edu.au 129.78.248.198
Message-ID: <347_38B72188.E79E2C6C@mail.usyd.edu.au>
Path: forums-1-dub!forums-1-dub!forums-master.sybase.com!forums.sybase.com
Xref: forums-1-dub sybase.public.easerver:27764
Article PK: 154582

I have been getting some weird memory leakage problems using jaguar cts
3.0 and the HTMLDW on windows NT.

On one machine ( dual p3 550, 128MB ) I am getting virtually no memory
loss. On a retrieve I can see that the private bytes for jagsrv
increases during processing but returns to the exact same value ( to the
last byte )

Now yesterday I created and installed two production machines each with
dual p3 550's and a gig of ram each. I now get larger memory increases
on an update and jaguar loses around 500K during a retrieve.

The installs are exactly the same code, same options on the objects. The
only difference was that I was accessing a totally different database on
a different server than development. I tried changing to an ODBC driver
to access the sybase database, but didnt seem to make much difference.
On average an update was losing between 600 and 1200K on these machines,
while development would never lose more than 500K.

If anyone could explain how this can happen it would be very much
appreciated.

Rgds
Adam


Download VCard asimmond.vcf


Dave Wolf [Sybase] Posted on 2000-02-26 15:13:29.0Z
Newsgroups: sybase.public.easerver
From: "Dave Wolf [Sybase]" <dwolf@sybase.com>
Subject: Re: Memory loss - Can someone explain this?
Date: Sat, 26 Feb 2000 10:13:29 -0500
Lines: 36
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 5.00.2919.6600
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2919.6600
NNTP-Posting-Host: vpn-eme-013.sybase.com 130.214.8.13
Message-ID: <347_gEIYQyGg$GA.274@forums.sybase.com>
References: <347_38B72188.E79E2C6C@mail.usyd.edu.au>
Path: forums-1-dub!forums-1-dub!forums-master.sybase.com!forums.sybase.com
Xref: forums-1-dub sybase.public.easerver:27746
Article PK: 154579

My guess is somewhere there are version mismatches of either PB, or
OpenClient or Jag. Id dig into that and see what you come up with thats
different.

Dave Wolf
Internet Applications Division

"Adam Simmonds" <asimmond@mail.usyd.edu.au> wrote in message
news:38B72188.E79E2C6C@mail.usyd.edu.au...
> I have been getting some weird memory leakage problems using jaguar cts
> 3.0 and the HTMLDW on windows NT.
>
> On one machine ( dual p3 550, 128MB ) I am getting virtually no memory
> loss. On a retrieve I can see that the private bytes for jagsrv
> increases during processing but returns to the exact same value ( to the
> last byte )
>
> Now yesterday I created and installed two production machines each with
> dual p3 550's and a gig of ram each. I now get larger memory increases
> on an update and jaguar loses around 500K during a retrieve.
>
> The installs are exactly the same code, same options on the objects. The
> only difference was that I was accessing a totally different database on
> a different server than development. I tried changing to an ODBC driver
> to access the sybase database, but didnt seem to make much difference.
> On average an update was losing between 600 and 1200K on these machines,
> while development would never lose more than 500K.
>
> If anyone could explain how this can happen it would be very much
> appreciated.
>
> Rgds
> Adam
>


Adam Simmonds Posted on 2000-02-27 01:14:02.0Z
Newsgroups: sybase.public.easerver
Date: Sun, 27 Feb 2000 12:14:02 +1100
From: Adam Simmonds <asimmond@mail.usyd.edu.au>
Organization: University of Syndey
X-Mailer: Mozilla 4.7 [en] (WinNT; I)
X-Accept-Language: en
MIME-Version: 1.0
Subject: Re: Memory loss - Can someone explain this?
Content-Type: multipart/mixed; boundary="------------DD6EB901EB4DBD191119CB6F"
Lines: 66
NNTP-Posting-Host: pc-70.sis.usyd.edu.au 129.78.248.198
Message-ID: <347_38B87A5A.731AC969@mail.usyd.edu.au>
References: <347_38B72188.E79E2C6C@mail.usyd.edu.au> <347_gEIYQyGg$GA.274@forums.sybase.com>
Path: forums-1-dub!forums-1-dub!forums-master.sybase.com!forums.sybase.com
Xref: forums-1-dub sybase.public.easerver:27732
Article PK: 154578

Yes its my own fault, and I was blaming poor old Jaguar :-)

I did some searching around and the PB versions were different. I had build
8017 on the machine that was okay with memory and 7017 on the one that was
leaking. Installed the latest EBF and it basically stops the leakge in its
tracks.

Was there something in the recent builds to address leakage in PB components
in jaguar?

Thanks Dave

Adam

"Dave Wolf [Sybase]" wrote:

> My guess is somewhere there are version mismatches of either PB, or
> OpenClient or Jag. Id dig into that and see what you come up with thats
> different.
>
> Dave Wolf
> Internet Applications Division
>
> "Adam Simmonds" <asimmond@mail.usyd.edu.au> wrote in message
> news:38B72188.E79E2C6C@mail.usyd.edu.au...
> > I have been getting some weird memory leakage problems using jaguar cts
> > 3.0 and the HTMLDW on windows NT.
> >
> > On one machine ( dual p3 550, 128MB ) I am getting virtually no memory
> > loss. On a retrieve I can see that the private bytes for jagsrv
> > increases during processing but returns to the exact same value ( to the
> > last byte )
> >
> > Now yesterday I created and installed two production machines each with
> > dual p3 550's and a gig of ram each. I now get larger memory increases
> > on an update and jaguar loses around 500K during a retrieve.
> >
> > The installs are exactly the same code, same options on the objects. The
> > only difference was that I was accessing a totally different database on
> > a different server than development. I tried changing to an ODBC driver
> > to access the sybase database, but didnt seem to make much difference.
> > On average an update was losing between 600 and 1200K on these machines,
> > while development would never lose more than 500K.
> >
> > If anyone could explain how this can happen it would be very much
> > appreciated.
> >
> > Rgds
> > Adam
> >


Download VCard asimmond.vcf


Jim O'Neil [Sybase] Posted on 2000-02-27 02:50:20.0Z
Newsgroups: sybase.public.easerver
Date: Sat, 26 Feb 2000 21:50:20 -0500
From: "Jim O'Neil [Sybase]" <joneil@sybase.com>
Organization: Sybase, Inc.
X-Mailer: Mozilla 4.7 [en]C-CCK-MCD (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
Subject: Re: Memory loss - Can someone explain this?
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Lines: 59
NNTP-Posting-Host: nomad6-48.sybase.com 157.133.176.48
Message-ID: <347_38B890EC.BEAB02A9@sybase.com>
References: <347_38B72188.E79E2C6C@mail.usyd.edu.au> <347_gEIYQyGg$GA.274@forums.sybase.com> <347_38B87A5A.731AC969@mail.usyd.edu.au>
Path: forums-1-dub!forums-1-dub!forums-master.sybase.com!forums.sybase.com
Xref: forums-1-dub sybase.public.easerver:27726
Article PK: 154577


Adam Simmonds wrote:

> Yes its my own fault, and I was blaming poor old Jaguar :-)
>
> I did some searching around and the PB versions were different. I had build
> 8017 on the machine that was okay with memory and 7017 on the one that was
> leaking. Installed the latest EBF and it basically stops the leakge in its
> tracks.
>
> Was there something in the recent builds to address leakage in PB components
> in jaguar?
>
> Thanks Dave
>
> Adam
>
> "Dave Wolf [Sybase]" wrote:
>
> > My guess is somewhere there are version mismatches of either PB, or
> > OpenClient or Jag. Id dig into that and see what you come up with thats
> > different.
> >
> > Dave Wolf
> > Internet Applications Division
> >
> > "Adam Simmonds" <asimmond@mail.usyd.edu.au> wrote in message
> > news:38B72188.E79E2C6C@mail.usyd.edu.au...
> > > I have been getting some weird memory leakage problems using jaguar cts
> > > 3.0 and the HTMLDW on windows NT.
> > >
> > > On one machine ( dual p3 550, 128MB ) I am getting virtually no memory
> > > loss. On a retrieve I can see that the private bytes for jagsrv
> > > increases during processing but returns to the exact same value ( to the
> > > last byte )
> > >
> > > Now yesterday I created and installed two production machines each with
> > > dual p3 550's and a gig of ram each. I now get larger memory increases
> > > on an update and jaguar loses around 500K during a retrieve.
> > >
> > > The installs are exactly the same code, same options on the objects. The
> > > only difference was that I was accessing a totally different database on
> > > a different server than development. I tried changing to an ODBC driver
> > > to access the sybase database, but didnt seem to make much difference.
> > > On average an update was losing between 600 and 1200K on these machines,
> > > while development would never lose more than 500K.
> > >
> > > If anyone could explain how this can happen it would be very much
> > > appreciated.
> > >
> > > Rgds
> > > Adam
> > >

There was a memory leak fixed with PB components that returns strings or blobs
(and the WebDW definitely returns strings) around the 7023 build or so, so that
could well explain what you are seeing.
--
Jim O'Neil
Senior Technical Support Engineer
Sybase, Inc


Adam Simmonds Posted on 2000-02-27 07:36:55.0Z
Newsgroups: sybase.public.easerver
Date: Sun, 27 Feb 2000 18:36:55 +1100
From: Adam Simmonds <asimmond@mail.usyd.edu.au>
Organization: University of Syndey
X-Mailer: Mozilla 4.7 [en] (WinNT; I)
X-Accept-Language: en
MIME-Version: 1.0
Subject: Re: Memory loss - Can someone explain this?
Content-Type: multipart/mixed; boundary="------------9288C52C76FCAA0F4A6266BE"
Lines: 83
NNTP-Posting-Host: pc-70.sis.usyd.edu.au 129.78.248.198
Message-ID: <347_38B8D417.4435222C@mail.usyd.edu.au>
References: <347_38B72188.E79E2C6C@mail.usyd.edu.au> <347_gEIYQyGg$GA.274@forums.sybase.com> <347_38B87A5A.731AC969@mail.usyd.edu.au> <347_38B890EC.BEAB02A9@sybase.com>
Path: forums-1-dub!forums-1-dub!forums-master.sybase.com!forums.sybase.com
Xref: forums-1-dub sybase.public.easerver:27723
Article PK: 159082

That would explain why one of my shared objects that returned the data in a
datastore as a string was blowing the memory away.

Thx
Adam

"Jim O'Neil [Sybase]" wrote:

> Adam Simmonds wrote:
>
> > Yes its my own fault, and I was blaming poor old Jaguar :-)
> >
> > I did some searching around and the PB versions were different. I had build
> > 8017 on the machine that was okay with memory and 7017 on the one that was
> > leaking. Installed the latest EBF and it basically stops the leakge in its
> > tracks.
> >
> > Was there something in the recent builds to address leakage in PB components
> > in jaguar?
> >
> > Thanks Dave
> >
> > Adam
> >
> > "Dave Wolf [Sybase]" wrote:
> >
> > > My guess is somewhere there are version mismatches of either PB, or
> > > OpenClient or Jag. Id dig into that and see what you come up with thats
> > > different.
> > >
> > > Dave Wolf
> > > Internet Applications Division
> > >
> > > "Adam Simmonds" <asimmond@mail.usyd.edu.au> wrote in message
> > > news:38B72188.E79E2C6C@mail.usyd.edu.au...
> > > > I have been getting some weird memory leakage problems using jaguar cts
> > > > 3.0 and the HTMLDW on windows NT.
> > > >
> > > > On one machine ( dual p3 550, 128MB ) I am getting virtually no memory
> > > > loss. On a retrieve I can see that the private bytes for jagsrv
> > > > increases during processing but returns to the exact same value ( to the
> > > > last byte )
> > > >
> > > > Now yesterday I created and installed two production machines each with
> > > > dual p3 550's and a gig of ram each. I now get larger memory increases
> > > > on an update and jaguar loses around 500K during a retrieve.
> > > >
> > > > The installs are exactly the same code, same options on the objects. The
> > > > only difference was that I was accessing a totally different database on
> > > > a different server than development. I tried changing to an ODBC driver
> > > > to access the sybase database, but didnt seem to make much difference.
> > > > On average an update was losing between 600 and 1200K on these machines,
> > > > while development would never lose more than 500K.
> > > >
> > > > If anyone could explain how this can happen it would be very much
> > > > appreciated.
> > > >
> > > > Rgds
> > > > Adam
> > > >
>
> There was a memory leak fixed with PB components that returns strings or blobs
> (and the WebDW definitely returns strings) around the 7023 build or so, so that
> could well explain what you are seeing.
> --
> Jim O'Neil
> Senior Technical Support Engineer
> Sybase, Inc


Download VCard asimmond.vcf


Dave Wolf [Sybase] Posted on 2000-02-27 01:25:04.0Z
Newsgroups: sybase.public.easerver
From: "Dave Wolf [Sybase]" <dwolf@sybase.com>
Subject: Re: Memory loss - Can someone explain this?
Date: Sat, 26 Feb 2000 20:25:04 -0500
Lines: 71
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 5.00.2919.6600
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2919.6600
NNTP-Posting-Host: vpn-eme-013.sybase.com 130.214.8.13
Message-ID: <347_A$XFBIMg$GA.202@forums.sybase.com>
References: <347_38B72188.E79E2C6C@mail.usyd.edu.au> <347_gEIYQyGg$GA.274@forums.sybase.com> <347_38B87A5A.731AC969@mail.usyd.edu.au>
Path: forums-1-dub!forums-1-dub!forums-master.sybase.com!forums.sybase.com
Xref: forums-1-dub sybase.public.easerver:27730
Article PK: 159089

Memory leaks are always a part of life. There have been known leaks in
previous builds, and we have fixed them as we have them identified. This
was luckily one we had already fixed :)

Dave Wolf
Internet Applications Division

"Adam Simmonds" <asimmond@mail.usyd.edu.au> wrote in message
news:38B87A5A.731AC969@mail.usyd.edu.au...
> Yes its my own fault, and I was blaming poor old Jaguar :-)
>
> I did some searching around and the PB versions were different. I had
build
> 8017 on the machine that was okay with memory and 7017 on the one that was
> leaking. Installed the latest EBF and it basically stops the leakge in its
> tracks.
>
> Was there something in the recent builds to address leakage in PB
components
> in jaguar?
>
> Thanks Dave
>
> Adam
>
> "Dave Wolf [Sybase]" wrote:
>
> > My guess is somewhere there are version mismatches of either PB, or
> > OpenClient or Jag. Id dig into that and see what you come up with thats
> > different.
> >
> > Dave Wolf
> > Internet Applications Division
> >
> > "Adam Simmonds" <asimmond@mail.usyd.edu.au> wrote in message
> > news:38B72188.E79E2C6C@mail.usyd.edu.au...
> > > I have been getting some weird memory leakage problems using jaguar
cts
> > > 3.0 and the HTMLDW on windows NT.
> > >
> > > On one machine ( dual p3 550, 128MB ) I am getting virtually no memory
> > > loss. On a retrieve I can see that the private bytes for jagsrv
> > > increases during processing but returns to the exact same value ( to
the
> > > last byte )
> > >
> > > Now yesterday I created and installed two production machines each
with
> > > dual p3 550's and a gig of ram each. I now get larger memory increases
> > > on an update and jaguar loses around 500K during a retrieve.
> > >
> > > The installs are exactly the same code, same options on the objects.
The
> > > only difference was that I was accessing a totally different database
on
> > > a different server than development. I tried changing to an ODBC
driver
> > > to access the sybase database, but didnt seem to make much difference.
> > > On average an update was losing between 600 and 1200K on these
machines,
> > > while development would never lose more than 500K.
> > >
> > > If anyone could explain how this can happen it would be very much
> > > appreciated.
> > >
> > > Rgds
> > > Adam
> > >
>


Adam Simmonds Posted on 2000-02-27 05:36:08.0Z
Newsgroups: sybase.public.easerver
Date: Sun, 27 Feb 2000 16:36:08 +1100
From: Adam Simmonds <asimmond@mail.usyd.edu.au>
Organization: University of Syndey
X-Mailer: Mozilla 4.7 [en] (WinNT; I)
X-Accept-Language: en
MIME-Version: 1.0
Subject: Re: Memory loss - Can someone explain this?
Content-Type: multipart/mixed; boundary="------------1C413A291D87C4BF178E1BF6"
Lines: 95
NNTP-Posting-Host: pc-70.sis.usyd.edu.au 129.78.248.198
Message-ID: <347_38B8B7C8.3F0EA098@mail.usyd.edu.au>
References: <347_38B72188.E79E2C6C@mail.usyd.edu.au> <347_gEIYQyGg$GA.274@forums.sybase.com> <347_38B87A5A.731AC969@mail.usyd.edu.au> <347_A$XFBIMg$GA.202@forums.sybase.com>
Path: forums-1-dub!forums-1-dub!forums-master.sybase.com!forums.sybase.com
Xref: forums-1-dub sybase.public.easerver:27724
Article PK: 159083

Thanks Dave, good work :-)

I am much much happier now, our UAT machine was dying very quickly.
My measurements identified around 2000-4000K of memory lost on an update and
around 1000K for a retrieve. It didnt take long for the server to start eating
virtual memory and then jaguar would hit the roof and fallover.

Now it runs beatifully, I have performed around 200 updates and memory is
remaining very constant ( except for when I make it create new instance via load
testing which would be expected )

So all is okay for now :-)

Thanks again

Adam

"Dave Wolf [Sybase]" wrote:

> Memory leaks are always a part of life. There have been known leaks in
> previous builds, and we have fixed them as we have them identified. This
> was luckily one we had already fixed :)
>
> Dave Wolf
> Internet Applications Division
>
> "Adam Simmonds" <asimmond@mail.usyd.edu.au> wrote in message
> news:38B87A5A.731AC969@mail.usyd.edu.au...
> > Yes its my own fault, and I was blaming poor old Jaguar :-)
> >
> > I did some searching around and the PB versions were different. I had
> build
> > 8017 on the machine that was okay with memory and 7017 on the one that was
> > leaking. Installed the latest EBF and it basically stops the leakge in its
> > tracks.
> >
> > Was there something in the recent builds to address leakage in PB
> components
> > in jaguar?
> >
> > Thanks Dave
> >
> > Adam
> >
> > "Dave Wolf [Sybase]" wrote:
> >
> > > My guess is somewhere there are version mismatches of either PB, or
> > > OpenClient or Jag. Id dig into that and see what you come up with thats
> > > different.
> > >
> > > Dave Wolf
> > > Internet Applications Division
> > >
> > > "Adam Simmonds" <asimmond@mail.usyd.edu.au> wrote in message
> > > news:38B72188.E79E2C6C@mail.usyd.edu.au...
> > > > I have been getting some weird memory leakage problems using jaguar
> cts
> > > > 3.0 and the HTMLDW on windows NT.
> > > >
> > > > On one machine ( dual p3 550, 128MB ) I am getting virtually no memory
> > > > loss. On a retrieve I can see that the private bytes for jagsrv
> > > > increases during processing but returns to the exact same value ( to
> the
> > > > last byte )
> > > >
> > > > Now yesterday I created and installed two production machines each
> with
> > > > dual p3 550's and a gig of ram each. I now get larger memory increases
> > > > on an update and jaguar loses around 500K during a retrieve.
> > > >
> > > > The installs are exactly the same code, same options on the objects.
> The
> > > > only difference was that I was accessing a totally different database
> on
> > > > a different server than development. I tried changing to an ODBC
> driver
> > > > to access the sybase database, but didnt seem to make much difference.
> > > > On average an update was losing between 600 and 1200K on these
> machines,
> > > > while development would never lose more than 500K.
> > > >
> > > > If anyone could explain how this can happen it would be very much
> > > > appreciated.
> > > >
> > > > Rgds
> > > > Adam
> > > >
> >


Download VCard asimmond.vcf