Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Problems creating replicas

0 views
Skip to first unread message

Pai Chung

unread,
Apr 9, 2002, 10:15:11 AM4/9/02
to
Hello everyone,

I have the following problem.
Each time I try to create a replica, it comes back right at the end of the
replica creation.
Error creating replica.
Couldn't create replica, because field is too small to accept value.

I tried several things, initially when I removed some defunct replicas it
worked, then it started failing again, next I tried compact/repair, recover
design master, shortening the retention period. None of these helped.

Anyone have any ideas?
Thanks,

Pai Chung
ch...@auditleverage.com

Richard Jarmain

unread,
Apr 10, 2002, 5:16:28 AM4/10/02
to
I wouldn't expect compacting, recovering etc to have any effect on this -
actually it sounds like you're more in need of a redesign of the data. I
suggest checking the replication white paper and your data.

"Pai Chung" <ch...@auditleverage.com> wrote in message
news:uZKdyD93BHA.2024@tkmsftngp02...

Pai Chung

unread,
Apr 10, 2002, 10:17:33 AM4/10/02
to
Hi Richard,

That is a good thought however this is as far as I can tell more a limit on
how many replicas I can create from an existing replica than anything else.
This problem has only occurred in locations where people have tried to spin
off more than 40 replicas.
The MsysReplicas table itself contains about 270 replicas most defunct,
however we have other locations with the same amount of items in the
MsysReplicas table that have no problem creating replicas, it looks more
like when you go over some "invisible limit" you run into this problem.

Pai Chung
ch...@auditleverage.com

"Richard Jarmain" <rich...@illuminaries.co.uk> wrote in message
news:Obr9H9G4BHA.2092@tkmsftngp07...

Michael (michka) Kaplan

unread,
Apr 11, 2002, 12:18:09 AM4/11/02
to
This is incorrect, there is no such limit at that number. There is a limit
of number of synchronizers, but that is 64,000.


--
MichKa

Michael Kaplan
Trigeminal Software, Inc. -- http://www.trigeminal.com/

International VB? -- http://www.i18nWithVB.com/
C++? MSLU -- http://msdn.microsoft.com/msdnmag/issues/01/10/


"Pai Chung" <ch...@auditleverage.com> wrote in message

news:O6qO2pJ4BHA.1344@tkmsftngp04...

Pai Chung

unread,
Apr 11, 2002, 9:35:40 AM4/11/02
to
Hi everyone,

Unfortunately the only similarity I see between the different scenarios that
are ongoing. Between those that can and can not create replicas, that is the
only observation I can make so far.
I guess what I wonder about is the following.
The creation of the replica fails AFTER the whole progress bar is completed.
What is it still trying to do at this point in time? Personally I figured it
was trying to update some system tables and that is where it ran into
problems.
This also made me think that there might be a limit on creating replicas
(although you can copy the replica using filecopy and create new replicas
that way but this is not a very good way of doing it)
Does anyone have anymore insights or thoughts on this?

Pai Chung
ch...@auditleverage.com

"Michael (michka) Kaplan" <forme...@nospam.trigeminal.spamless.com> wrote
in message news:eQ03z#Q4BHA.2284@tkmsftngp02...

Richard Jarmain

unread,
Apr 12, 2002, 1:03:41 PM4/12/02
to
Hi,

Michka is of course right, you can be certain that it is not a limit on the
number of replicas which is causing your problem. I'm afraid I can only
offer something similar to before: in my experience, it isn't likely to be a
problem with the system tables - especially with the relatively low numbers
you're dealing with (eg 40 and 270 - imagine the system tables for a company
with 1,000 travelling salesmen!), it's much more likely to be in the data
model somewhere.
Best of luck....


"Pai Chung" <ch...@auditleverage.com> wrote in message

news:uu07M3V4BHA.2836@tkmsftngp05...

Gary Hall

unread,
Apr 18, 2002, 9:43:48 AM4/18/02
to
Gentleman,

I think Richard guessed right when he said the problem has
to be in the system tables. I've had this same problem
for months. I have been watching this newsgroup for
awhile and I see this problem come up occasionally. No
one really has an answer. I can tell you that even with
all the objects deleted, the error still occurs. This
does seem to point to the system tables.

Gary

>> > > > > IÍ{ wÀ "]_",Ç

Michael (michka) Kaplan

unread,
Apr 18, 2002, 1:38:57 PM4/18/02
to
You did not read Richard's responses very carefully:

"in my experience, it isn't likely to be a
problem with the system tables"

and in the earlier mail:

"I wouldn't expect compacting, recovering etc to
have any effect on this - actually it sounds like
you're more in need of a redesign of the data."


--
MichKa

Michael Kaplan
Trigeminal Software, Inc. -- http://www.trigeminal.com/

"Gary Hall" <gary...@bolling.af.mil> wrote in message
news:3cf501c1e6df$112bc060$a5e62ecf@tkmsftngxa07...
Gentleman,

Gary

>> > > > > Iヘ{掫タ "]_",ヌ

Gary Hall

unread,
May 21, 2002, 11:29:22 AM5/21/02
to
Michael,

I make the following observation:

After I delete all the objects in the database, the
problem still occurs. This does not seem to point to the
data model, unless I'm missing something here.

Gary

>>> > > Hi RicharÍ{ wÀ -¾7L Â7,Ûâ ì d,

Michael (michka) Kaplan

unread,
May 21, 2002, 12:31:06 PM5/21/02
to
You are missing the point. A replica contains meta-information about the database itself in the system tables. Even deleting all objects will not delete the information (it has to be kept so that older clients can still sync). It is still more likely a design issue. Because well-designed replicas never seem to report this problem? MichKa Michael Kaplan Trigeminal Software, Inc. -- http://www.trigeminal.com/ International VB? -- http://www.i18nWithVB.com/ C++? MSLU -- http://msdn.microsoft.com/msdnmag/issues/01/10/ "Gary Hall" <gary...@bolling.af.mil> wrote in message news:51b701c200dc$4856d310$36ef2ecf@tkmsftngxa12... Michael, I make the following observation: After I delete all the objects in the database, the problem still occurs. This does not seem to point to the data model, unless I'm missing something here. Gary >-----Original Message----- >You did not read Richard's responses very carefully: >"in my experience, it isn't likely to be a >problem with the system tables" >and in the earlier mail: >"I wouldn't expect compacting, recovering etc to >have any effect on this - actually it sounds like >you're more in need of a redesign of the data." >MichKa >Michael Kaplan >Trigeminal Software, Inc. -- http://www.trigeminal.com/ >International VB? -- http://www.i18nWithVB.com/ >C++? MSLU -- http://msdn.microsoft.com/msdnmag/issues/01/10/ >"Gary Hall" <gary...@bolling.af.mil> wrote in message >news:3cf501c1e6df$112bc060$a5e62ecf@tkmsftngxa07... >Gentleman, >I think Richard guessed right when he said the problem has >to be in the system tables. I've had this same problem >for months. I have been watching this newsgroup for >awhile and I see this problem come up occasionally. No >one really has an answer. I can tell you that even with >all the objects deleted, the error still occurs. This >does seem to point to the system tables. >Gary >>-----Original Message----- >>Hi, >>Michka is of course right, you can be certain that it is >not a limit on the >>number of replicas which is causing your problem. I'm >afraid I can only >>offer something similar to before: in my experience, it >isn't likely to be a >>problem with the system tables - especially with the >relatively low numbers >>you're dealing with (eg 40 and 270 - imagine the system >tables for a company >>with 1,000 travelling salesmen!), it's much more likely >to be in the data >>model somewhere. >>Best of luck.... >>"Pai Chung" <ch...@auditleverage.com> wrote in message >>news:uu07M3V4BHA.2836@tkmsftngp05... >>> Hi everyone, >>> Unfortunately the only similarity I see between the >different scenarios >>that >>> are ongoing. Between those that can and can not create >replicas, that is >>the >>> only observation I can make so far. >>> I guess what I wonder about is the following. >>> The creation of the replica fails AFTER the whole >progress bar is >>completed. >>> What is it still trying to do at this point in time? >Personally I figured >>it >>> was trying to update some system tables and that is >where it ran into >>> problems. >>> This also made me think that there might be a limit on >creating replicas >>> (although you can copy the replica using filecopy and >create new replicas >>> that way but this is not a very good way of doing it) >>> Does anyone have anymore insights or thoughts on this? >>> Pai Chung >>> ch...@auditleverage.com >>> "Michael (michka) Kaplan" ><forme...@nospam.trigeminal.spamless.com> >>wrote >>> in message news:eQ03z#Q4BHA.2284@tkmsftngp02... >>> > This is incorrect, there is no such limit at that >number. There is a >>limit >>> > of number of synchronizers, but that is 64,000. >>> > >>> > >>> > -- >>> > MichKa >>> > >>> > Michael Kaplan >>> > Trigeminal Software, Inc. -- >http://www.trigeminal.com/ >>> > >>> > International VB? -- http://www.i18nWithVB.com/ >>> > C++? MSLU -- >http://msdn.microsoft.com/msdnmag/issues/01/10/ >>> > >>> > >>> > "Pai Chung" <ch...@auditleverage.com> wrote in message >>> > news:O6qO2pJ4BHA.1344@tkmsftngp04... >>> > > Hi Richar >>> > > >>> > > That is a good thought however this is as far as I >can tell more a >>limit >>> > on >>> > > how many replicas I can create from an existing >replica than anything >>> > else. >>> > > This problem has only occurred in locations where >people have tried to >>> > spi

Gary Hall

unread,
May 21, 2002, 1:15:56 PM5/21/02
to
MichKa,

I'll be the first to agree with you that I am missing the
point. What I don't understand is your definition of a
well designed replica. I'm no newbie at this stuff....My
design master is a simple back-end with tables only,
well....just one module containing one function. I
typically synchronize my design master with what I call
my "Hub Replica" which sits on a network drive. I use
this hub to create replicas for my users. Do you have any
other guidance?

Gary

>-----Original Message-----
>You are missing the point.
>
>A replica contains meta-information about the database
itself in the system
>tables. Even deleting all objects will not delete the
information (it has to
>be kept so that older clients can still sync).
>
>It is still more likely a design issue. Because well-
designed replicas never
>seem to report this problem?
>
>

>--
>MichKa
>
>Michael Kaplan
>Trigeminal Software, Inc. -- http://www.trigeminal.com/
>
>International VB? -- http://www.i18nWithVB.com/
>C++? MSLU --
http://msdn.microsoft.com/msdnmag/issues/01/10/
>
>

>"Gary Hall" <gary...@bolling.af.mil> wrote in message
>news:51b701c200dc$4856d310$36ef2ecf@tkmsftngxa12...
>Michael,
>
>I make the following observation:
>
>After I delete all the objects in the database, the
>problem still occurs. This does not seem to point to the
>data model, unless I'm missing something here.
>
>Gary
>>-----Original Message-----
>>You did not read Richard's responses very carefully:
>>
>>"in my experience, it isn't likely to be a
>>problem with the system tables"
>>
>>and in the earlier mail:
>>
>>"I wouldn't expect compacting, recovering etc to
>>have any effect on this - actually it sounds like
>>you're more in need of a redesign of the data."
>>
>>

>>--
>>MichKa
>>
>>Michael Kaplan
>>Trigeminal Software, Inc. -- http://www.trigeminal.com/
>>
>>International VB? -- http://www.i18nWithVB.com/
>>C++? MSLU --
>http://msdn.microsoft.com/msdnmag/issues/01/10/
>>
>>
>>

>>>> was trying to update Í{ wÀ f | YKìåYKÌS% ì some

Michael (michka) Kaplan

unread,
May 21, 2002, 1:36:24 PM5/21/02
to
Since you have not really given additional info, not sure how anyone can
diagnose the problem for you. In a casual glance you have about 10 messages
between December and now on variuous topics, none of which would aid here,
either.

This is not really a request for more info specifically, just an
observation. :-)


--
MichKa

Michael Kaplan
Trigeminal Software, Inc. -- http://www.trigeminal.com/


"Gary Hall" <gary...@bolling.af.mil> wrote in message

news:51bd01c200eb$2ba71540$9ee62ecf@tkmsftngxa05...
MichKa,

Gary

>>>> was trying to update ヘ{掫タ f | YK�藝KフS% � some

Gary Hall

unread,
May 21, 2002, 2:08:09 PM5/21/02
to
MichKa,

Thanks for hanging tough on this one. Ok, here goes.
I'll try and give you all the relevant info I can muster.

I have a "Design Master" containing 30 or so tables (and
one module). I use this "Design Master" to transfer
design changes to my "Hub Replica" which sits on the
server. I synchronize the "Design Master" with the hub
once a week just so it contains up to date data. I also
clear up any conflicts that I find, then remove the
conflict tables after I'm done. Every so often, when I
open the "Hub Replica" to create a new replica on a local
machine, I get the error message "The field is too small
to accept the amount of data you attempted to add. Try
inserting or pasting less data." When I try recreating
the "Hub Replica" from my "Design Master", I get the same
error message. I usually resolve this problem by finding
some other replica in the set that can still create
replicas...and bebuilding my "Hub Replica" and "Design
Master." Inevitably, a few days will pass and the problem
will re-occur. I'm trying to figure out what is causing
the error. I guess I could spend an entire day making
another replica set but I'm leary of this happening
again.....although, I have about 8 other backends that I
maintain using the same repllication scheme and I have no
other problems with them.

Gary

>>>one really has aÍ{ wÀ Q ¼?´L´jµL # ì. n answer. I

>>>>> was trying to update Í{ wÀ f | YKìåYKÌS% ì some


>system tables and that is
>>>where it ran into
>>>>> problems.
>>>>> This also made me think that there might be a limit
on
>>>creating replicas
>>>>> (although you can copy the replica using filecopy and
>>>create new replicas
>>>>> that way but this is not a very good way of doing it
>
>

>.
>

Michael (michka) Kaplan

unread,
May 21, 2002, 2:23:31 PM5/21/02
to
How about you just freeze the design and not make so many changes to it?


--
MichKa

Michael Kaplan
Trigeminal Software, Inc. -- http://www.trigeminal.com/


"Gary Hall" <gary...@bolling.af.mil> wrote in message

news:5ae301c200f2$76ca72e0$9be62ecf@tkmsftngxa03...

Vitalijus J. Karalius

unread,
May 22, 2002, 1:08:53 AM5/22/02
to
But as a purely practical matter, since everthing is essentially "the same"
between the backends, and the others have no issues, that *something* has
changed this set. Although one always wants the answer to *why*, the real
world solution is bite the bullet, go replica to regular and recreate a new
set. Does not take that long even on large backends with lots of tables -
even more than 30. Otherwise, will you ever be "sure" about your data? We
use the design master rather than the hub to create any new replicas since
there is intentionally less interaction between the DM and the set, and we
can clean up any issues prior to a final sync with the DM, followed by
create new replica.


"Gary Hall" <gary...@bolling.af.mil> wrote in message

news:5ae301c200f2$76ca72e0$9be62ecf@tkmsftngxa03...

0 new messages