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.

Replication does not survive server restart

18 posts in Replication Last posting was on 2008-12-09 15:19:16.0Z
Tom Arleth Posted on 2008-09-29 12:15:01.0Z
From: "Tom Arleth" <ta@ascott.dk>
Newsgroups: Advantage.Replication
Subject: Replication does not survive server restart
Date: Mon, 29 Sep 2008 14:15:01 +0200
Lines: 25
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.5512
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.5579
X-RFC2646: Format=Flowed; Original
NNTP-Posting-Host: 62.242.32.53
Message-ID: <48e0c5d9@solutions.advantagedatabase.com>
X-Trace: 29 Sep 2008 06:11:05 -0700, 62.242.32.53
Path: solutions.advantagedatabase.com!solutions.advantagedatabase.com!62.242.32.53
Xref: solutions.advantagedatabase.com Advantage.Replication:316
Article PK: 1134171

Hi,

Some of our customers have quite frequent server restarts (power failures
where the UPS does not suffice). On several of these occasions the
replication does not survive the restart. That is there is no entries in the
replication queue and nothing is getting transmitted to the receiver.
Nothing that I can see seems to be wrong execpt it does not work.
At pressent I have to define a new publication and subscription to get
things started, which means that I also have to make a new syncronisation of
tha databases to get the "lost" entries "transmitted", which means that I
have to shut their system down for maintainence for a few hours to make sure
that I have syncronised versions when I restart the replication. Not very
nice neither for me nor for them.
Is there something that has to or can be done so that the replication starts
automatically when the server restarts?

The server is a Win2003 server with advantage 8.1 (dot 26 I think)

--
--
Venlig hilsen
Tom Arleth
Ascott Software Danmark A/S


Edgar Sherman Posted on 2008-09-29 15:03:41.0Z
Date: Mon, 29 Sep 2008 09:03:41 -0600
From: Edgar Sherman <no@email.com>
User-Agent: Thunderbird 2.0.0.16 (Windows/20080708)
MIME-Version: 1.0
Newsgroups: Advantage.Replication
Subject: Re: Replication does not survive server restart
References: <48e0c5d9@solutions.advantagedatabase.com>
In-Reply-To: <48e0c5d9@solutions.advantagedatabase.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
NNTP-Posting-Host: 10.24.34.164
Message-ID: <48e0ecc6@solutions.advantagedatabase.com>
X-Trace: 29 Sep 2008 08:57:10 -0700, 10.24.34.164
Lines: 37
Path: solutions.advantagedatabase.com!solutions.advantagedatabase.com!10.24.34.164
Xref: solutions.advantagedatabase.com Advantage.Replication:317
Article PK: 1134172

Tom,

I am not sure I quite understand the issue. Here are two things I would
check.

First, if the source server is restarted, you must log into the
dictionary for replication to begin again.

Second, Make sure that when you log into the dictionary that you login
as "REMOTE" server and NOT "LOCAL" server. LOCAL server does not have
replication so this may be giving you trouble if the application
automatically scales back to "LOCAL" server.

If there is nothing in the replication queue, try modifying a record and
make sure it gets to the queue.

Edgar

Tom Arleth wrote:
> Hi,
>
> Some of our customers have quite frequent server restarts (power failures
> where the UPS does not suffice). On several of these occasions the
> replication does not survive the restart. That is there is no entries in the
> replication queue and nothing is getting transmitted to the receiver.
> Nothing that I can see seems to be wrong execpt it does not work.
> At pressent I have to define a new publication and subscription to get
> things started, which means that I also have to make a new syncronisation of
> tha databases to get the "lost" entries "transmitted", which means that I
> have to shut their system down for maintainence for a few hours to make sure
> that I have syncronised versions when I restart the replication. Not very
> nice neither for me nor for them.
> Is there something that has to or can be done so that the replication starts
> automatically when the server restarts?
>
> The server is a Win2003 server with advantage 8.1 (dot 26 I think)
>


Tom Arleth Posted on 2008-09-30 09:51:14.0Z
From: "Tom Arleth" <ta@ascott.dk>
Newsgroups: Advantage.Replication
References: <48e0c5d9@solutions.advantagedatabase.com> <48e0ecc6@solutions.advantagedatabase.com>
Subject: Re: Replication does not survive server restart
Date: Tue, 30 Sep 2008 11:51:14 +0200
Lines: 53
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.5512
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.5579
X-RFC2646: Format=Flowed; Response
NNTP-Posting-Host: 62.242.32.53
Message-ID: <48e1f5ac@solutions.advantagedatabase.com>
X-Trace: 30 Sep 2008 03:47:24 -0700, 62.242.32.53
Path: solutions.advantagedatabase.com!solutions.advantagedatabase.com!62.242.32.53
Xref: solutions.advantagedatabase.com Advantage.Replication:318
Article PK: 1134173

Hi Edgar,

All connections are dictionary connections with remote server.

When I see the error the customer may have been running on the system for
days without generating any entries. I cant generate any entries either - I
try to do so by manually editing a record with the Arc but nothing happens.

--
Venlig hilsen
Tom Arleth
Ascott Software Danmark A/S
"Edgar Sherman" <no@email.com> skrev i en meddelelse
news:48e0ecc6@solutions.advantagedatabase.com...

> Tom,
>
> I am not sure I quite understand the issue. Here are two things I would
> check.
>
> First, if the source server is restarted, you must log into the dictionary
> for replication to begin again.
>
> Second, Make sure that when you log into the dictionary that you login as
> "REMOTE" server and NOT "LOCAL" server. LOCAL server does not have
> replication so this may be giving you trouble if the application
> automatically scales back to "LOCAL" server.
>
> If there is nothing in the replication queue, try modifying a record and
> make sure it gets to the queue.
>
> Edgar
>
> Tom Arleth wrote:
>> Hi,
>>
>> Some of our customers have quite frequent server restarts (power failures
>> where the UPS does not suffice). On several of these occasions the
>> replication does not survive the restart. That is there is no entries in
>> the replication queue and nothing is getting transmitted to the receiver.
>> Nothing that I can see seems to be wrong execpt it does not work.
>> At pressent I have to define a new publication and subscription to get
>> things started, which means that I also have to make a new syncronisation
>> of tha databases to get the "lost" entries "transmitted", which means
>> that I have to shut their system down for maintainence for a few hours to
>> make sure that I have syncronised versions when I restart the
>> replication. Not very nice neither for me nor for them.
>> Is there something that has to or can be done so that the replication
>> starts automatically when the server restarts?
>>
>> The server is a Win2003 server with advantage 8.1 (dot 26 I think)
>>


Tom Arleth Posted on 2008-11-13 13:42:31.0Z
From: "Tom Arleth" <ta@ascott.dk>
Newsgroups: Advantage.Replication
References: <48e0c5d9@solutions.advantagedatabase.com> <48e0ecc6@solutions.advantagedatabase.com> <48e1f5ac@solutions.advantagedatabase.com>
Subject: Re: Replication does not survive server restart
Date: Thu, 13 Nov 2008 14:42:31 +0100
Lines: 73
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.5512
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.5579
X-RFC2646: Format=Flowed; Response
NNTP-Posting-Host: 62.242.32.53
Message-ID: <491c2e0e@solutions.advantagedatabase.com>
X-Trace: 13 Nov 2008 06:39:26 -0700, 62.242.32.53
Path: solutions.advantagedatabase.com!solutions.advantagedatabase.com!62.242.32.53
Xref: solutions.advantagedatabase.com Advantage.Replication:319
Article PK: 1134174

Hi,

Any thougths on why replication has to be started/recreated manually when
the server has been restarted?
It is really a major sequrity risk that the database is not replicated just
because the server has been restarted! (There is a "real" backup as well but
it is in the same server and the tape is only changed on a weekly basis)


--
Venlig hilsen
Tom Arleth
Ascott Software Danmark A/S

"Tom Arleth" <ta@ascott.dk> skrev i en meddelelse
news:48e1f5ac@solutions.advantagedatabase.com...

> Hi Edgar,
>
> All connections are dictionary connections with remote server.
>
> When I see the error the customer may have been running on the system for
> days without generating any entries. I cant generate any entries either -
> I try to do so by manually editing a record with the Arc but nothing
> happens.
>
> --
> Venlig hilsen
> Tom Arleth
> Ascott Software Danmark A/S
> "Edgar Sherman" <no@email.com> skrev i en meddelelse
> news:48e0ecc6@solutions.advantagedatabase.com...
>> Tom,
>>
>> I am not sure I quite understand the issue. Here are two things I would
>> check.
>>
>> First, if the source server is restarted, you must log into the
>> dictionary for replication to begin again.
>>
>> Second, Make sure that when you log into the dictionary that you login
>> as "REMOTE" server and NOT "LOCAL" server. LOCAL server does not have
>> replication so this may be giving you trouble if the application
>> automatically scales back to "LOCAL" server.
>>
>> If there is nothing in the replication queue, try modifying a record and
>> make sure it gets to the queue.
>>
>> Edgar
>>
>> Tom Arleth wrote:
>>> Hi,
>>>
>>> Some of our customers have quite frequent server restarts (power
>>> failures where the UPS does not suffice). On several of these occasions
>>> the replication does not survive the restart. That is there is no
>>> entries in the replication queue and nothing is getting transmitted to
>>> the receiver. Nothing that I can see seems to be wrong execpt it does
>>> not work.
>>> At pressent I have to define a new publication and subscription to get
>>> things started, which means that I also have to make a new
>>> syncronisation of tha databases to get the "lost" entries "transmitted",
>>> which means that I have to shut their system down for maintainence for a
>>> few hours to make sure that I have syncronised versions when I restart
>>> the replication. Not very nice neither for me nor for them.
>>> Is there something that has to or can be done so that the replication
>>> starts automatically when the server restarts?
>>>
>>> The server is a Win2003 server with advantage 8.1 (dot 26 I think)
>>>
>
>


Peter Funk (ADS) Posted on 2008-11-13 15:46:05.0Z
Date: Thu, 13 Nov 2008 15:46:05 +0000 (UTC)
Message-ID: <864d0bcb15fe68cb138366566495@devzone.advantagedatabase.com>
From: Peter Funk (ADS) <pfunk@nospam.com>
Subject: Re: Replication does not survive server restart
Newsgroups: Advantage.Replication
References: <491c2e0e@solutions.advantagedatabase.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Content-Type: text/plain; charset=iso-8859-1; format=flowed
X-Newsreader: JetBrains Omea Pro 1098.1
NNTP-Posting-Host: 10.24.38.185
X-Trace: 13 Nov 2008 08:39:40 -0700, 10.24.38.185
Lines: 18
Path: solutions.advantagedatabase.com!solutions.advantagedatabase.com!10.24.38.185
Xref: solutions.advantagedatabase.com Advantage.Replication:320
Article PK: 1134175

Hello Tom,
I'm stumped as to why your server isn't replicating after being restarted.
The very first thing you should do is check the ads_err log on the server
(should be in c:\) to see if there are any errors related to replication.
I would expect to see some erorrs there when the server attempts to place
replication updates in the queue. Do you have the replication setup to ignore
replication errors? If so, then you might not see any errors in the error
log or entries in the queue. Otherwise, if there are no errors in the log
or entries in the queue, then it sounds like the server doesn't think replication
is setup. That would lead me to believe that there is something wrong with
your data dictionary after the restart. In that case, it would help a lot
if you could send us a copy of your dictionary in that state.

Regards,
Peter Funk
Advantage R&D


Tom Arleth Posted on 2008-11-14 07:59:39.0Z
From: "Tom Arleth" <ta@ascott.dk>
Newsgroups: Advantage.Replication
References: <491c2e0e@solutions.advantagedatabase.com> <864d0bcb15fe68cb138366566495@devzone.advantagedatabase.com>
Subject: Re: Replication does not survive server restart
Date: Fri, 14 Nov 2008 08:59:39 +0100
Lines: 38
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.5512
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.5579
X-RFC2646: Format=Flowed; Response
NNTP-Posting-Host: 62.242.32.53
Message-ID: <491d2f2e@solutions.advantagedatabase.com>
X-Trace: 14 Nov 2008 00:56:30 -0700, 62.242.32.53
Path: solutions.advantagedatabase.com!solutions.advantagedatabase.com!62.242.32.53
Xref: solutions.advantagedatabase.com Advantage.Replication:321
Article PK: 1134176

Hi Peter,

There is not any errors related to the replication in the ads_err.adt at the
time of the latest restart. There are some errors that I cant seem to find
in the help file (7154 and 10054) and some errors regarding the queue
(7039+7008) arriving before the last restart (and before the last
replication was performed).

Will a simple filecopy of the ".add" file suffice (server is in use 24/7) or
maybe the ".add" file from the live backup?

--
Venlig hilsen
Tom Arleth
Ascott Software Danmark A/S
"Peter Funk (ADS)" <pfunk@nospam.com> skrev i en meddelelse
news:864d0bcb15fe68cb138366566495@devzone.advantagedatabase.com...

> Hello Tom,
> I'm stumped as to why your server isn't replicating after being restarted.
> The very first thing you should do is check the ads_err log on the server
> (should be in c:\) to see if there are any errors related to replication.
> I would expect to see some erorrs there when the server attempts to place
> replication updates in the queue. Do you have the replication setup to
> ignore replication errors? If so, then you might not see any errors in
> the error log or entries in the queue. Otherwise, if there are no errors
> in the log or entries in the queue, then it sounds like the server doesn't
> think replication is setup. That would lead me to believe that there is
> something wrong with your data dictionary after the restart. In that
> case, it would help a lot if you could send us a copy of your dictionary
> in that state.
>
> Regards,
> Peter Funk
> Advantage R&D
>
>


Peter Funk (ADS) Posted on 2008-11-14 21:10:05.0Z
Date: Fri, 14 Nov 2008 21:10:05 +0000 (UTC)
Message-ID: <864d0bcb160278cb1479d3b36c24@devzone.advantagedatabase.com>
From: Peter Funk (ADS) <pfunk@nospam.com>
Subject: Re: Replication does not survive server restart
Newsgroups: Advantage.Replication
References: <491d2f2e@solutions.advantagedatabase.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Content-Type: text/plain; charset=iso-8859-1; format=flowed
X-Newsreader: JetBrains Omea Pro 1098.1
NNTP-Posting-Host: 10.24.38.185
X-Trace: 14 Nov 2008 14:03:40 -0700, 10.24.38.185
Lines: 33
Path: solutions.advantagedatabase.com!solutions.advantagedatabase.com!10.24.38.185
Xref: solutions.advantagedatabase.com Advantage.Replication:322
Article PK: 1134177

Hello Tom,
The 7154 error indicates that ADS automatically rebuilt an index because
it wasn't closed properly (almost certainly from the unexpected restart).
When ADS caches index information it will flag the index as corrupt until
all updates have been written to the file on disk. If the corruption flag
is still set when ADS opens the index, it will automatically rebuild it.
This is a new error code in 8.1 that missed the help documentation (it is
in the 9.1 help file).

The 10054 error is a windows networking error described as "An existing connection
was forcibly closed by the remote host." It is a common error code to see
if a client loses its network connection to ADS.

I played around with some test dictionaries and tried to re-create what is
happening to your server. I set the corruption flag in the replication queue
index, reconnected to the dictionary and updated the replicated table and
subsequent replication updates worked as expected. ADS rebuilt the queue
index and logged the 7154 error like you saw in your error log.

So there must be something else going on that is preventing your server from
replicating. If you can, please send me your .add and .am files of the dictionary
from just after a server restart. You can do a simple filecopy to get them,
but if that doesn't work then a copy of the backup files would be fine too.
If you can't get the dictionary files from just after a restart, I might
be able to re-create that situation using a copy of the normal files. Also
include your error log files if you can. Please send them to advantage@ianywhere.com
(attn: Peter Funk).

Regards,
Peter Funk
Advantage R&D


Tom Arleth Posted on 2008-11-17 14:54:45.0Z
From: "Tom Arleth" <ta@ascott.dk>
Newsgroups: Advantage.Replication
References: <491d2f2e@solutions.advantagedatabase.com> <864d0bcb160278cb1479d3b36c24@devzone.advantagedatabase.com>
Subject: Re: Replication does not survive server restart
Date: Mon, 17 Nov 2008 15:54:45 +0100
Lines: 55
X-Newsreader: Microsoft Outlook Express 6.00.2900.5512
X-MIMEOLE: Produced By Microsoft MimeOLE V6.00.2900.5579
X-RFC2646: Format=Flowed; Response
NNTP-Posting-Host: 62.242.32.53
Message-ID: <49218452@solutions.advantagedatabase.com>
X-Trace: 17 Nov 2008 07:48:50 -0700, 62.242.32.53
Path: solutions.advantagedatabase.com!solutions.advantagedatabase.com!62.242.32.53
Xref: solutions.advantagedatabase.com Advantage.Replication:323
Article PK: 1134178

Hi Peter,

I sent them earlier today but forgot to mention that the last replikated
rekord was created 27/19/2008 15:25:21 and that restarts of the server (and
thus presumably shortly before also shut-dows) were performed at the
following days:
2008-10-27 15:59:13: ADS Service started shortly (~10 seks) after.
2008-10-28 11:37:19: ADS Service started shortly after.
2008-10-28 16:23:45: ADS Service started shortly after.
2008-10-28 21:06:27: ADS Service started shortly after.
2008-11-03 17:19:27: ADS Service started shortly after.

--
Kind Regards
Tom Arleth
Ascott Software Danmark A/S


"Peter Funk (ADS)" <pfunk@nospam.com> skrev i en meddelelse
news:864d0bcb160278cb1479d3b36c24@devzone.advantagedatabase.com...

> Hello Tom,
> The 7154 error indicates that ADS automatically rebuilt an index because
> it wasn't closed properly (almost certainly from the unexpected restart).
> When ADS caches index information it will flag the index as corrupt until
> all updates have been written to the file on disk. If the corruption flag
> is still set when ADS opens the index, it will automatically rebuild it.
> This is a new error code in 8.1 that missed the help documentation (it is
> in the 9.1 help file).
>
> The 10054 error is a windows networking error described as "An existing
> connection was forcibly closed by the remote host." It is a common error
> code to see if a client loses its network connection to ADS.
>
> I played around with some test dictionaries and tried to re-create what is
> happening to your server. I set the corruption flag in the replication
> queue index, reconnected to the dictionary and updated the replicated
> table and subsequent replication updates worked as expected. ADS rebuilt
> the queue index and logged the 7154 error like you saw in your error log.
>
> So there must be something else going on that is preventing your server
> from replicating. If you can, please send me your .add and .am files of
> the dictionary from just after a server restart. You can do a simple
> filecopy to get them, but if that doesn't work then a copy of the backup
> files would be fine too. If you can't get the dictionary files from just
> after a restart, I might be able to re-create that situation using a copy
> of the normal files. Also include your error log files if you can.
> Please send them to advantage@ianywhere.com (attn: Peter Funk).
>
> Regards,
> Peter Funk
> Advantage R&D
>
>


Peter Funk (ADS) Posted on 2008-11-20 21:12:35.0Z
Date: Thu, 20 Nov 2008 21:12:35 +0000 (UTC)
Message-ID: <864d0bcb1618c8cb19312c1ad7ab@devzone.advantagedatabase.com>
From: Peter Funk (ADS) <pfunk@nospam.com>
Subject: Re: Replication does not survive server restart
Newsgroups: Advantage.Replication
References: <49218452@solutions.advantagedatabase.com>
MIME-Version: 1.0
Content-Transfer-Encoding: Quoted-Printable
Content-Type: multipart/mixed; boundary="--++Omea_Parts_Splitter.788792905299362"
X-Newsreader: JetBrains Omea Pro 1098.1
NNTP-Posting-Host: 10.24.38.185
X-Trace: 20 Nov 2008 14:06:09 -0700, 10.24.38.185
Lines: 41
Path: solutions.advantagedatabase.com!solutions.advantagedatabase.com!10.24.38.185
Xref: solutions.advantagedatabase.com Advantage.Replication:324
Article PK: 1134179

Hello Tom,
I've been looking at your error log and dictionary files and I'm still unsure what the problem could be. The server is trying to open the replication queue but it fails with a 7039 (File already open) error. This error means the table is already open, but with different open options (sharing mode, locking mode, etc). Do you open the replication queue directly at any time? If you had the queue open directly, then the server might not be able to open it and place updates in it.

I think the best thing for us to do is to get a dump of the server's internal memory at the time of the 7039 error. All you need to do is set a few registry keys and restart Advantage. I attached a .reg file with the keys in it and all you need to do is double-click the .reg file on the server and it will merge the keys into the Advantage configuration. Here are the settings in case the .reg file doesn't get through:

Windows Registry Editor Version 5.00

[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Advantage\Configuration]
"minidumpfile"="tblutil.c"
"minidumperror"=dword:00001b7f
"minidumpline"=dword:00000581


These settings will cause Advantage to create a dump file when it encounters the 7039 error on a specific line of code. It will only create one dump a day so you don't need to worry about it creating too many dump files. The dump will be created at the root of c:\ and it'll be called adsdump-YYYYMMDD-HHMMSS.dmp.gz. Once you have one or more of these please send them in to me. I can give you instructions on uploading them to our FTP server if you like.

Regards,
Peter Funk
Advantage R&D


Tom Arleth Posted on 2008-11-21 07:29:35.0Z
From: "Tom Arleth" <ta@ascott.dk>
Newsgroups: Advantage.Replication
References: <49218452@solutions.advantagedatabase.com> <864d0bcb1618c8cb19312c1ad7ab@devzone.advantagedatabase.com>
Subject: Re: Replication does not survive server restart
Date: Fri, 21 Nov 2008 08:29:35 +0100
Lines: 53
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.5512
X-RFC2646: Format=Flowed; Original
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.5579
NNTP-Posting-Host: 62.242.32.53
Message-ID: <49266200@solutions.advantagedatabase.com>
X-Trace: 21 Nov 2008 00:23:44 -0700, 62.242.32.53
Path: solutions.advantagedatabase.com!solutions.advantagedatabase.com!62.242.32.53
Xref: solutions.advantagedatabase.com Advantage.Replication:325
Article PK: 1134180

Hej Peter,

That makes some sence. We do acces the queue every 10 minutes from a
seperate service in order to monitor whether the replication is still
running (is it empty or with a new item on top then everything is fine -
otherwhise a problem may exist). Maybe that service opens the queue in a
less that optimal way. I'll install the reg changes and inspect our code
(can you tell me whitch open options shoul not have been used?).

--
Venlig hilsen
Tom Arleth
Ascott Software Danmark A/S

"Peter Funk (ADS)" <pfunk@nospam.com> skrev i en meddelelse
news:864d0bcb1618c8cb19312c1ad7ab@devzone.advantagedatabase.com...
Hello Tom,
I've been looking at your error log and dictionary files and I'm still
unsure what the problem could be. The server is trying to open the
replication queue but it fails with a 7039 (File already open) error. This
error means the table is already open, but with different open options
(sharing mode, locking mode, etc). Do you open the replication queue
directly at any time? If you had the queue open directly, then the server
might not be able to open it and place updates in it.

I think the best thing for us to do is to get a dump of the server's
internal memory at the time of the 7039 error. All you need to do is set a
few registry keys and restart Advantage. I attached a .reg file with the
keys in it and all you need to do is double-click the .reg file on the
server and it will merge the keys into the Advantage configuration. Here
are the settings in case the .reg file doesn't get through:

Windows Registry Editor Version 5.00

[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Advantage\Configuration]
"minidumpfile"="tblutil.c"
"minidumperror"=dword:00001b7f
"minidumpline"=dword:00000581


These settings will cause Advantage to create a dump file when it encounters
the 7039 error on a specific line of code. It will only create one dump a
day so you don't need to worry about it creating too many dump files. The
dump will be created at the root of c:\ and it'll be called
adsdump-YYYYMMDD-HHMMSS.dmp.gz. Once you have one or more of these please
send them in to me. I can give you instructions on uploading them to our
FTP server if you like.

Regards,
Peter Funk
Advantage R&D


Peter Funk (ADS) Posted on 2008-11-21 15:11:15.0Z
Date: Fri, 21 Nov 2008 15:11:15 +0000 (UTC)
Message-ID: <864d0bcb161ad8cb19c7dbe76cfc@devzone.advantagedatabase.com>
From: Peter Funk (ADS) <pfunk@nospam.com>
Subject: Re: Replication does not survive server restart
Newsgroups: Advantage.Replication
References: <49266200@solutions.advantagedatabase.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Content-Type: text/plain; charset=iso-8859-1; format=flowed
X-Newsreader: JetBrains Omea Pro 1098.1
NNTP-Posting-Host: 10.24.38.185
X-Trace: 21 Nov 2008 08:04:49 -0700, 10.24.38.185
Lines: 16
Path: solutions.advantagedatabase.com!solutions.advantagedatabase.com!10.24.38.185
Xref: solutions.advantagedatabase.com Advantage.Replication:326
Article PK: 1134181

Hello Tom,
The 7039 error is a mismatch in sharing modes. Either the table was opened
exclusively and then failed to be opened again, or it was opened shared and
failed to be opened exclusively later. I know the replication code will
open the queue shared, so I would recommend that anytime you open it in your
code it should be shared as well. One caveat is that if the table wasn't
closed properly (as in an unexpected server restart), ADS will briefly have
the queue open exclusively in order to rebuild the index the first time the
queue is opened. My original hunch was that this was causing your problems,
but now I'm not so sure. The error logs seem to tell a different story.

Regards,
Peter Funk
Advantage R&D


Tom Arleth Posted on 2008-11-24 09:24:52.0Z
From: "Tom Arleth" <ta@ascott.dk>
Newsgroups: Advantage.Replication
References: <49266200@solutions.advantagedatabase.com> <864d0bcb161ad8cb19c7dbe76cfc@devzone.advantagedatabase.com>
Subject: Re: Replication does not survive server restart
Date: Mon, 24 Nov 2008 10:24:52 +0100
Lines: 541
MIME-Version: 1.0
Content-Type: multipart/mixed; boundary="----=_NextPart_000_0062_01C94E1E.E388D270"
X-Newsreader: Microsoft Outlook Express 6.00.2900.5512
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.5579
NNTP-Posting-Host: 62.242.32.53
Message-ID: <492a7189@solutions.advantagedatabase.com>
X-Trace: 24 Nov 2008 02:19:05 -0700, 62.242.32.53
Path: solutions.advantagedatabase.com!solutions.advantagedatabase.com!62.242.32.53
Xref: solutions.advantagedatabase.com Advantage.Replication:328
Article PK: 1134183

Hi Peter,
 
The code that opens the code looks like this (AscOpenConnection just sets the rigth username and password):
  AdsConn := TAdsConnection.Create(nil);
  try
    AdsConn.Name := 'RepMonService' + IntToStr(GetTickCount);
    AdsConn.ConnectPath := Self.FDictionary;
    AdsConn.LoginPrompt := False;
    AdsConn.Username := Self.FAdsUser;
    AdsConn.AdsServerTypes := [stADS_REMOTE];
 
    AscOpenConnection(AdsConn);
    try
      RepQueue.AdsConnection := AdsConn;
      RepQueue.SQL.Text := 'SELECT * FROM ' + Self.FRepQueueName + ' ORDER BY EntryID';
      RepQueue.Open;
      if RepQueue.Bof and RepQueue.Eof then
However I dont think that this is directly involved in the problem because it was not executed on the troubled restarts. I have taken a look in the servers event log, and it appears (see attached image) that my Replication Monitoring Service starts up before the ADS service and hence cant connect to the database. It then crashes (with the error: Error 6097: Bad IP address specified in the connection path or in the ADS.INI file. axServerConnect) and stops ~4 seks before the ADS starts.
 
 
 

-- 
Kind Regards
Tom Arleth
Ascott Software Danmark A/S
 
 
"Peter Funk (ADS)" <pfunk@nospam.com> skrev i en meddelelse news:864d0bcb161ad8cb19c7dbe76cfc@devzone.advantagedatabase.com...
> Hello Tom,
> The 7039 error is a mismatch in sharing modes.  Either the table was opened
> exclusively and then failed to be opened again, or it was opened shared and
> failed to be opened exclusively later.  I know the replication code will
> open the queue shared, so I would recommend that anytime you open it in your
> code it should be shared as well.  One caveat is that if the table wasn't
> closed properly (as in an unexpected server restart), ADS will briefly have
> the queue open exclusively in order to rebuild the index the first time the
> queue is opened.  My original hunch was that this was causing your problems,
> but now I'm not so sure.  The error logs seem to tell a different story.
>
> Regards,
> Peter Funk
> Advantage R&D
>
>

Replikering eventlog tasiilaq.png

Peter Funk (ADS) Posted on 2008-11-21 21:34:29.0Z
Date: Fri, 21 Nov 2008 21:34:29 +0000 (UTC)
Message-ID: <864d0bcb161cb8cb19fd6555b0cc@devzone.advantagedatabase.com>
From: Peter Funk (ADS) <pfunk@nospam.com>
Subject: Re: Replication does not survive server restart
Newsgroups: Advantage.Replication
References: <49266200@solutions.advantagedatabase.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Content-Type: text/plain; charset=iso-8859-1; format=flowed
X-Newsreader: JetBrains Omea Pro 1098.1
NNTP-Posting-Host: 10.24.38.185
X-Trace: 21 Nov 2008 14:28:02 -0700, 10.24.38.185
Lines: 18
Path: solutions.advantagedatabase.com!solutions.advantagedatabase.com!10.24.38.185
Xref: solutions.advantagedatabase.com Advantage.Replication:327
Article PK: 1134182

Some more thoughts about this issue:
We're suspcious that the index of the dictionary (the .ai file) might be
getting corrupt. If possible we would really like to get a copy of the dictionary
when the replication isn't working. Even if the index wasn't corrupt, if
we could see the problem occuring here we could almost certainly tell what
the problem is.

One suggestion for the next time the problem occurs is to delete the .ai
file of the dictionary. No users can be connected to it at the time, but
the next time it is opened, ADS will automatically re-create it. If the
problem is with the dictionary index, then forcing ADS to rebuild it should
fix it.

Regards,
Peter Funk
Advantage R&D


Tom Arleth Posted on 2008-11-24 09:40:29.0Z
From: "Tom Arleth" <ta@ascott.dk>
Newsgroups: Advantage.Replication
References: <49266200@solutions.advantagedatabase.com> <864d0bcb161cb8cb19fd6555b0cc@devzone.advantagedatabase.com>
Subject: Re: Replication does not survive server restart
Date: Mon, 24 Nov 2008 10:40:29 +0100
Lines: 40
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.5512
X-RFC2646: Format=Flowed; Response
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.5579
NNTP-Posting-Host: 62.242.32.53
Message-ID: <492a7532@solutions.advantagedatabase.com>
X-Trace: 24 Nov 2008 02:34:42 -0700, 62.242.32.53
Path: solutions.advantagedatabase.com!solutions.advantagedatabase.com!62.242.32.53
Xref: solutions.advantagedatabase.com Advantage.Replication:329
Article PK: 1134184

Hi Peter,

The backup you got last week was made at a time where the replication did
not work - actually it still does not as I have abstained from "fixing" it.
No changes in the database are finding their way to the replication queue,
which is what the situation has been like 27/19/2008 15:25:21.

I plan on fixing it in the near future (I need a confirmation from their
administrator before I start throwing off their users from tha database
:-) ). I could try doing it by deleting the .ai file - that is easier than
deleting and recreating the replication setup which is what I have been
doing so far when I repair the replication.

--
Kind regards
Tom Arleth
Ascott Software Danmark A/S

"Peter Funk (ADS)" <pfunk@nospam.com> skrev i en meddelelse
news:864d0bcb161cb8cb19fd6555b0cc@devzone.advantagedatabase.com...

> Some more thoughts about this issue:
> We're suspcious that the index of the dictionary (the .ai file) might be
> getting corrupt. If possible we would really like to get a copy of the
> dictionary when the replication isn't working. Even if the index wasn't
> corrupt, if we could see the problem occuring here we could almost
> certainly tell what the problem is.
>
> One suggestion for the next time the problem occurs is to delete the .ai
> file of the dictionary. No users can be connected to it at the time, but
> the next time it is opened, ADS will automatically re-create it. If the
> problem is with the dictionary index, then forcing ADS to rebuild it
> should fix it.
>
> Regards,
> Peter Funk
> Advantage R&D
>
>


Mark Wilkins Posted on 2008-11-24 21:24:59.0Z
From: "Mark Wilkins" <mark@no.email>
Newsgroups: Advantage.Replication
References: <49266200@solutions.advantagedatabase.com> <864d0bcb161cb8cb19fd6555b0cc@devzone.advantagedatabase.com> <492a7532@solutions.advantagedatabase.com>
Subject: Re: Replication does not survive server restart
Date: Mon, 24 Nov 2008 14:24:59 -0700
Lines: 36
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.3138
X-RFC2646: Format=Flowed; Response
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.3350
NNTP-Posting-Host: 10.24.38.228
Message-ID: <492b1a28@solutions.advantagedatabase.com>
X-Trace: 24 Nov 2008 14:18:32 -0700, 10.24.38.228
Path: solutions.advantagedatabase.com!solutions.advantagedatabase.com!10.24.38.228
Xref: solutions.advantagedatabase.com Advantage.Replication:330
Article PK: 1134186

Hi Tom,

While it is in this state, it would be useful if you could make a copy (just
a normal file copy) of the .add, .am, and .ai files (same as last time but
with the .ai file included this time). And possibly more useful might be to
generate a dump file of the server's state. The simplest way to do that is
probably through Advantage Data Architect. Open the management utility and
connect to the source server. Then press <ctrl>F9. This causes Advantage
to write a dump file (all the current state) to a file named
adsdump-<datetime>.dmp.gz. It will be placed in the same location as the
ads_err.* files. If you can send us that file, that could be very useful in
understanding the problem. Note that when you create the dump file,
Advantage freezes all threads during the write of that file, so it brings
things to standstill (the total time depends on how much memory is
allocated).

Mark Wilkins
Advantage R&D

"Tom Arleth" <ta@ascott.dk> wrote in message
news:492a7532@solutions.advantagedatabase.com...
> Hi Peter,
>
> The backup you got last week was made at a time where the replication did
> not work - actually it still does not as I have abstained from "fixing"
> it. No changes in the database are finding their way to the replication
> queue, which is what the situation has been like 27/19/2008 15:25:21.
>
> I plan on fixing it in the near future (I need a confirmation from their
> administrator before I start throwing off their users from tha database
> :-) ). I could try doing it by deleting the .ai file - that is easier than
> deleting and recreating the replication setup which is what I have been
> doing so far when I repair the replication.
>


Mark Wilkins Posted on 2008-12-04 17:00:09.0Z
From: "Mark Wilkins" <mark@no.email>
Newsgroups: Advantage.Replication
References: <48e0c5d9@solutions.advantagedatabase.com>
Subject: Re: Replication does not survive server restart
Date: Thu, 4 Dec 2008 10:00:09 -0700
Lines: 50
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.3138
X-RFC2646: Format=Flowed; Response
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.3350
NNTP-Posting-Host: 10.24.38.228
Message-ID: <49380b14@solutions.advantagedatabase.com>
X-Trace: 4 Dec 2008 09:53:40 -0700, 10.24.38.228
Path: solutions.advantagedatabase.com!solutions.advantagedatabase.com!10.24.38.228
Xref: solutions.advantagedatabase.com Advantage.Replication:331
Article PK: 1134185

We were able to track this down to a permissions issue in the server. The
problem occurs if the following conditions are met:

- Data dictionary has user rights checking turned on.
- Data dictionary allows anonymous logins (logins without a user name).
- The first connection to the data dictionary after a server restart is
the anonymous user.

In this situation, the server does not correctly populate the cache of
replication information for the dictionary. When tables are opened, the
server does not recognize them as being replicated.

The best way to ensure this problem does not occur is to disable anonymous
logins. You can right click on the dictionary name in the connection
repository of Advantage Data Architect and choose Properties. Then click on
the Security tab and select the "Logins Required" option.

This issue will be fixed in the next service update of 8.1x and 9.1x.

Mark Wilkins
Advantage R&D

"Tom Arleth" <ta@ascott.dk> wrote in message
news:48e0c5d9@solutions.advantagedatabase.com...
> Hi,
>
> Some of our customers have quite frequent server restarts (power failures
> where the UPS does not suffice). On several of these occasions the
> replication does not survive the restart. That is there is no entries in
> the replication queue and nothing is getting transmitted to the receiver.
> Nothing that I can see seems to be wrong execpt it does not work.
> At pressent I have to define a new publication and subscription to get
> things started, which means that I also have to make a new syncronisation
> of tha databases to get the "lost" entries "transmitted", which means that
> I have to shut their system down for maintainence for a few hours to make
> sure that I have syncronised versions when I restart the replication. Not
> very nice neither for me nor for them.
> Is there something that has to or can be done so that the replication
> starts automatically when the server restarts?
>
> The server is a Win2003 server with advantage 8.1 (dot 26 I think)
>
> --
> --
> Venlig hilsen
> Tom Arleth
> Ascott Software Danmark A/S
>


Tom Arleth Posted on 2008-12-09 07:46:35.0Z
From: "Tom Arleth" <ta@ascott.dk>
Newsgroups: Advantage.Replication
References: <48e0c5d9@solutions.advantagedatabase.com> <49380b14@solutions.advantagedatabase.com>
Subject: Re: Replication does not survive server restart
Date: Tue, 9 Dec 2008 08:46:35 +0100
Lines: 30
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.5512
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.5579
X-RFC2646: Format=Flowed; Response
NNTP-Posting-Host: 62.242.32.53
Message-ID: <493e2110@solutions.advantagedatabase.com>
X-Trace: 9 Dec 2008 00:41:04 -0700, 62.242.32.53
Path: solutions.advantagedatabase.com!solutions.advantagedatabase.com!62.242.32.53
Xref: solutions.advantagedatabase.com Advantage.Replication:332
Article PK: 1134187

Hi Mark,

"Mark Wilkins" <mark@no.email> skrev i en meddelelse
news:49380b14@solutions.advantagedatabase.com...
> We were able to track this down to a permissions issue in the server.

Zap

> The best way to ensure this problem does not occur is to disable anonymous
> logins.

Unfortunately that is not a viable option for us as we depend on the
anonymous login as the first login.

> This issue will be fixed in the next service update of 8.1x and 9.1x.

Is there any ETA for the 8.1 serviceupdate (or a chance for an EBF)?

>
> Mark Wilkins
> Advantage R&D
>


--
Kind regards
Tom Arleth
Ascott Software Danmark A/S


Mark Wilkins Posted on 2008-12-09 15:19:16.0Z
From: "Mark Wilkins" <mark@no.email>
Newsgroups: Advantage.Replication
References: <48e0c5d9@solutions.advantagedatabase.com> <49380b14@solutions.advantagedatabase.com> <493e2110@solutions.advantagedatabase.com>
Subject: Re: Replication does not survive server restart
Date: Tue, 9 Dec 2008 08:19:16 -0700
Lines: 30
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.5512
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.5579
X-RFC2646: Format=Flowed; Response
NNTP-Posting-Host: 10.24.38.228
Message-ID: <493e8aef@solutions.advantagedatabase.com>
X-Trace: 9 Dec 2008 08:12:47 -0700, 10.24.38.228
Path: solutions.advantagedatabase.com!solutions.advantagedatabase.com!10.24.38.228
Xref: solutions.advantagedatabase.com Advantage.Replication:333
Article PK: 1134188


"Tom Arleth" <ta@ascott.dk> wrote in message
news:493e2110@solutions.advantagedatabase.com...
> Hi Mark,
>
>> The best way to ensure this problem does not occur is to disable
>> anonymous logins.
>
> Unfortunately that is not a viable option for us as we depend on the
> anonymous login as the first login.

Does it have to be the anonymous login? Is it possible to set up a user
with no rights and no password? Aside from needing to provide a user name,
the functionality of such a user would be equivalent to the anonymous user.
If you were able to do that, then replication would not encounter the
problem.

>
>> This issue will be fixed in the next service update of 8.1x and 9.1x.
>
> Is there any ETA for the 8.1 serviceupdate (or a chance for an EBF)?

Since we just released an 8.1 update, the next one will probably be a while.
If it is not possible to create a user as described above, then I can build
and test an update.

Mark Wilkins
Advantage R&D