Client Connection/Disconnection!

Discussion:

(too old to reply)

John Vai

2005-04-09 15:52:05 UTC

Dear Experts! it's me again!
I have lost the old newsgoups on my system and still synchronizing the old
messages.

I would like to address my case once again for all:
I have a Win2k dual processor server SCSII hard disk that runs the dbsrv50
5.5.5 build 2817.
I have 8 users connecting to the server some win98 and some XP
The problem was that the engine is crashing or disconnecting all the clients
for a reason.
After upgrading from 5.5.4 to 5.5.5 (advice of Breck) with runing the
maintenance patch the system stopped 10 days later instead of stopping every
2-3 days later.
I have saved the log files of the engine and found out that sometimes the
reason of the crash/stop is causing assertion failed errors before the stop.

Here are the last 3 errors that I have gotten for the last 3 stops.
2 errors before the upgrade:
1)
11:37:30 *** ERROR *** Assertion failed: 50213
11:37:30 Page number on page does not match page requested
2)
13:48:32 Starting checkpoint: Wed Mar 23 2005 13:48
13:48:32 Fatal error: disk full C:\BIMDb\POS.db

the first stop(error) after the upgrade to 5.5.5
11:39:47 *** ERROR *** Assertion failed: 50204
11:39:47 When adding the first read lock to a cache page, a write lock was
found

can anyone? maybe tell me what's happening.
after the upgrade the system lasted for more than before.
or can you tell me HOW to prevent such assertions or what are causing such
assertions???

P.S. sometimes the engine was stopping without any error message.

I guess that the engine is stopping because of an assertion! but how to
prevent it????

PLEASE I AM IN A DEEP TROUBLE and only left are you experts!

John Vai

Paul Horan[TeamSybase]

2005-04-09 17:09:09 UTC

Permalink

Here's a clue:
Fatal error: disk full C:\BIMDb\POS.db

You're also looking at an unload/reload to correct the "Page number on page..." error.
--
Paul Horan[TeamSybase]

Post by John Vai
Dear Experts! it's me again!
I have lost the old newsgoups on my system and still synchronizing the old messages.
I have a Win2k dual processor server SCSII hard disk that runs the dbsrv50 5.5.5 build 2817.
I have 8 users connecting to the server some win98 and some XP
The problem was that the engine is crashing or disconnecting all the clients for a reason.
After upgrading from 5.5.4 to 5.5.5 (advice of Breck) with runing the maintenance patch the system stopped 10 days
later instead of stopping every 2-3 days later.
I have saved the log files of the engine and found out that sometimes the reason of the crash/stop is causing
assertion failed errors before the stop.
Here are the last 3 errors that I have gotten for the last 3 stops.
1)
11:37:30 *** ERROR *** Assertion failed: 50213
11:37:30 Page number on page does not match page requested
2)
13:48:32 Starting checkpoint: Wed Mar 23 2005 13:48
13:48:32 Fatal error: disk full C:\BIMDb\POS.db
the first stop(error) after the upgrade to 5.5.5
11:39:47 *** ERROR *** Assertion failed: 50204
11:39:47 When adding the first read lock to a cache page, a write lock was found
can anyone? maybe tell me what's happening.
after the upgrade the system lasted for more than before.
or can you tell me HOW to prevent such assertions or what are causing such assertions???
P.S. sometimes the engine was stopping without any error message.
I guess that the engine is stopping because of an assertion! but how to prevent it????
PLEASE I AM IN A DEEP TROUBLE and only left are you experts!
John Vai

Breck Carter [TeamSybase]

2005-04-09 19:45:53 UTC

Permalink

Assertion 50204 is very rare... I can find only one other report from
back in 1998; there was no reported cause or cure, but the result was
a corrupted database that had to be restored from backup.

(I hope you are making regular backups)

Can you check the disk drive for errors?

Also, do a check for RAM errors... sometimes problems with RAM *look*
like disk errors.

Is the disk drive almost full? As Paul pointed out, you did have a
"disk full" error before you upgraded.

Breck

On 9 Apr 2005 08:52:05 -0700, "John Vai" <john_vai[at]hotmail[dot]com>

Post by John Vai
Dear Experts! it's me again!
I have lost the old newsgoups on my system and still synchronizing the old
messages.
I have a Win2k dual processor server SCSII hard disk that runs the dbsrv50
5.5.5 build 2817.
I have 8 users connecting to the server some win98 and some XP
The problem was that the engine is crashing or disconnecting all the clients
for a reason.
After upgrading from 5.5.4 to 5.5.5 (advice of Breck) with runing the
maintenance patch the system stopped 10 days later instead of stopping every
2-3 days later.
I have saved the log files of the engine and found out that sometimes the
reason of the crash/stop is causing assertion failed errors before the stop.
Here are the last 3 errors that I have gotten for the last 3 stops.
1)
11:37:30 *** ERROR *** Assertion failed: 50213
11:37:30 Page number on page does not match page requested
2)
13:48:32 Starting checkpoint: Wed Mar 23 2005 13:48
13:48:32 Fatal error: disk full C:\BIMDb\POS.db
the first stop(error) after the upgrade to 5.5.5
11:39:47 *** ERROR *** Assertion failed: 50204
11:39:47 When adding the first read lock to a cache page, a write lock was
found
can anyone? maybe tell me what's happening.
after the upgrade the system lasted for more than before.
or can you tell me HOW to prevent such assertions or what are causing such
assertions???
P.S. sometimes the engine was stopping without any error message.
I guess that the engine is stopping because of an assertion! but how to
prevent it????
PLEASE I AM IN A DEEP TROUBLE and only left are you experts!
John Vai

--
SQL Anywhere Studio 9 Developer's Guide
Buy the book: http://www.amazon.com/exec/obidos/ASIN/1556225067/risingroad-20
***@risingroad.com
RisingRoad SQL Anywhere and MobiLink Professional Services
www.risingroad.com

John Vai

2005-04-09 23:42:54 UTC

Permalink

Maybe i can say that I am not that expert in Sybase but I do know some
important techniques.
First of all if my database is valid( no errros while validating the
database) does it mean that it could be that my database is corrupted
somehow?
Second, I didn't get what paul said: 'You're also looking at an
unload/reload to correct the "Page number on page..." error.' [Please paul i
know that you want me to understand what you mean but i don't get it, maybe
because i lack some techniques). if you could clarify.
Third, I tried in the past to run the engine on another (one
processor-win2k) server the result was the same, engine stoping and
disconnecting the clients so I guess that the servers are ok and the RAM is
ok too.

Guys can it be that i am doing something wrong in my Application which is
causing assertions?? but in other places with the same consitions no
problems at all.

Please ask me anything you want me to try but don't let me work it all
alone; what wonders me is that I cannot trace it down or know the real
problem.

Post by Breck Carter [TeamSybase]
Assertion 50204 is very rare... I can find only one other report from
back in 1998; there was no reported cause or cure, but the result was
a corrupted database that had to be restored from backup.
(I hope you are making regular backups)
Can you check the disk drive for errors?
Also, do a check for RAM errors... sometimes problems with RAM *look*
like disk errors.
Is the disk drive almost full? As Paul pointed out, you did have a
"disk full" error before you upgraded.
Breck
On 9 Apr 2005 08:52:05 -0700, "John Vai" <john_vai[at]hotmail[dot]com>

Post by John Vai
Dear Experts! it's me again!
I have lost the old newsgoups on my system and still synchronizing the old
messages.
I have a Win2k dual processor server SCSII hard disk that runs the dbsrv50
5.5.5 build 2817.
I have 8 users connecting to the server some win98 and some XP
The problem was that the engine is crashing or disconnecting all the clients
for a reason.
After upgrading from 5.5.4 to 5.5.5 (advice of Breck) with runing the
maintenance patch the system stopped 10 days later instead of stopping every
2-3 days later.
I have saved the log files of the engine and found out that sometimes the
reason of the crash/stop is causing assertion failed errors before the stop.
Here are the last 3 errors that I have gotten for the last 3 stops.
1)
11:37:30 *** ERROR *** Assertion failed: 50213
11:37:30 Page number on page does not match page requested
2)
13:48:32 Starting checkpoint: Wed Mar 23 2005 13:48
13:48:32 Fatal error: disk full C:\BIMDb\POS.db
the first stop(error) after the upgrade to 5.5.5
11:39:47 *** ERROR *** Assertion failed: 50204
11:39:47 When adding the first read lock to a cache page, a write lock was
found
can anyone? maybe tell me what's happening.
after the upgrade the system lasted for more than before.
or can you tell me HOW to prevent such assertions or what are causing such
assertions???
P.S. sometimes the engine was stopping without any error message.
I guess that the engine is stopping because of an assertion! but how to
prevent it????
PLEASE I AM IN A DEEP TROUBLE and only left are you experts!
John Vai

--
SQL Anywhere Studio 9 Developer's Guide
http://www.amazon.com/exec/obidos/ASIN/1556225067/risingroad-20
RisingRoad SQL Anywhere and MobiLink Professional Services
www.risingroad.com

Breck Carter [TeamSybase]

2005-04-10 15:36:48 UTC

Permalink

Please understand that this newsgroup is operated by volunteers, and
it is not a substitute for calling tech support. That is what you
should probably do if you want somebody to dedicate themselves to your
problem; see http://www.sybase.com/contactus/support#tech

If you run dbunload to unload the database into a reload.sql file and
*.dat files, dbinit to create a new *.db file, and isql to run
reload.sql, that will eliminate any database corruption. It may come
back, however, if there is a bug in the software.

Please undertand that version 5.5 is very very old, and any new bugs
will *not* be fixed.

Breck

On 9 Apr 2005 16:42:54 -0700, "John Vai" <john_vai[at]hotmail[dot]com>

Post by John Vai
Maybe i can say that I am not that expert in Sybase but I do know some
important techniques.
First of all if my database is valid( no errros while validating the
database) does it mean that it could be that my database is corrupted
somehow?
Second, I didn't get what paul said: 'You're also looking at an
unload/reload to correct the "Page number on page..." error.' [Please paul i
know that you want me to understand what you mean but i don't get it, maybe
because i lack some techniques). if you could clarify.
Third, I tried in the past to run the engine on another (one
processor-win2k) server the result was the same, engine stoping and
disconnecting the clients so I guess that the servers are ok and the RAM is
ok too.
Guys can it be that i am doing something wrong in my Application which is
causing assertions?? but in other places with the same consitions no
problems at all.
Please ask me anything you want me to try but don't let me work it all
alone; what wonders me is that I cannot trace it down or know the real
problem.

Post by John Vai
Dear Experts! it's me again!
I have lost the old newsgoups on my system and still synchronizing the old
messages.
I have a Win2k dual processor server SCSII hard disk that runs the dbsrv50
5.5.5 build 2817.
I have 8 users connecting to the server some win98 and some XP
The problem was that the engine is crashing or disconnecting all the clients
for a reason.
After upgrading from 5.5.4 to 5.5.5 (advice of Breck) with runing the
maintenance patch the system stopped 10 days later instead of stopping every
2-3 days later.
I have saved the log files of the engine and found out that sometimes the
reason of the crash/stop is causing assertion failed errors before the stop.
Here are the last 3 errors that I have gotten for the last 3 stops.
1)
11:37:30 *** ERROR *** Assertion failed: 50213
11:37:30 Page number on page does not match page requested
2)
13:48:32 Starting checkpoint: Wed Mar 23 2005 13:48
13:48:32 Fatal error: disk full C:\BIMDb\POS.db
the first stop(error) after the upgrade to 5.5.5
11:39:47 *** ERROR *** Assertion failed: 50204
11:39:47 When adding the first read lock to a cache page, a write lock was
found
can anyone? maybe tell me what's happening.
after the upgrade the system lasted for more than before.
or can you tell me HOW to prevent such assertions or what are causing such
assertions???
P.S. sometimes the engine was stopping without any error message.
I guess that the engine is stopping because of an assertion! but how to
prevent it????
PLEASE I AM IN A DEEP TROUBLE and only left are you experts!
John Vai

--
SQL Anywhere Studio 9 Developer's Guide
http://www.amazon.com/exec/obidos/ASIN/1556225067/risingroad-20
RisingRoad SQL Anywhere and MobiLink Professional Services
www.risingroad.com

John Vai

2005-04-10 22:23:39 UTC

Permalink

I really appreciate your help and Thank all of you for giving me some of
your time.
The problem is that we do not have any *Tech support* in Lebanon nor Middle
East.
I will try to contact somehow the tech support that you provided after
trying to recreate the database.
I will be submitting any constructive news that I might face in the future.

Thanks again.
John Vai

Post by Breck Carter [TeamSybase]
Please understand that this newsgroup is operated by volunteers, and
it is not a substitute for calling tech support. That is what you
should probably do if you want somebody to dedicate themselves to your
problem; see http://www.sybase.com/contactus/support#tech
If you run dbunload to unload the database into a reload.sql file and
*.dat files, dbinit to create a new *.db file, and isql to run
reload.sql, that will eliminate any database corruption. It may come
back, however, if there is a bug in the software.
Please undertand that version 5.5 is very very old, and any new bugs
will *not* be fixed.
Breck
On 9 Apr 2005 16:42:54 -0700, "John Vai" <john_vai[at]hotmail[dot]com>

Post by John Vai
Dear Experts! it's me again!
I have lost the old newsgoups on my system and still synchronizing the old
messages.
I have a Win2k dual processor server SCSII hard disk that runs the dbsrv50
5.5.5 build 2817.
I have 8 users connecting to the server some win98 and some XP
The problem was that the engine is crashing or disconnecting all the clients
for a reason.
After upgrading from 5.5.4 to 5.5.5 (advice of Breck) with runing the
maintenance patch the system stopped 10 days later instead of stopping every
2-3 days later.
I have saved the log files of the engine and found out that sometimes the
reason of the crash/stop is causing assertion failed errors before the stop.
Here are the last 3 errors that I have gotten for the last 3 stops.
1)
11:37:30 *** ERROR *** Assertion failed: 50213
11:37:30 Page number on page does not match page requested
2)
13:48:32 Starting checkpoint: Wed Mar 23 2005 13:48
13:48:32 Fatal error: disk full C:\BIMDb\POS.db
the first stop(error) after the upgrade to 5.5.5
11:39:47 *** ERROR *** Assertion failed: 50204
11:39:47 When adding the first read lock to a cache page, a write lock was
found
can anyone? maybe tell me what's happening.
after the upgrade the system lasted for more than before.
or can you tell me HOW to prevent such assertions or what are causing such
assertions???
P.S. sometimes the engine was stopping without any error message.
I guess that the engine is stopping because of an assertion! but how to
prevent it????
PLEASE I AM IN A DEEP TROUBLE and only left are you experts!
John Vai

--
SQL Anywhere Studio 9 Developer's Guide
http://www.amazon.com/exec/obidos/ASIN/1556225067/risingroad-20
RisingRoad SQL Anywhere and MobiLink Professional Services
www.risingroad.com

John Vai

2005-04-11 18:00:30 UTC

Permalink

For more information:
I got today a new assertion which is: 51901
Page for requested record not a table page or record not present in page.

I tried to search in the newsgroups about this specific assertion; what i
got is that they suggested to increase the page size to 4k.
I had some other posts from experts about problem causing the LONG VARCHAR
in tables. I have plenty of them in my table (why i don't know). I have
stored procedures that run other procedures too; and I have offered a -c
500MB cash which some have asked to decrease the cash to lower values for
some cases (Which i don't either do know).
I will be doing the following:
I will try to add page size to 4k, decrease the 500MB cash to 50MB (database
size is 30MB) and trace it down again.

I wanted to post this in case you would still try to find out.

What do you think???

Paul Horan[TeamSybase]

2005-04-11 20:00:54 UTC

Permalink

Post by John Vai
I got today a new assertion which is: 51901
Page for requested record not a table page or record not present in page.
I tried to search in the newsgroups about this specific assertion; what i got is that they suggested to increase the
page size to 4k.
I had some other posts from experts about problem causing the LONG VARCHAR in tables. I have plenty of them in my
table (why i don't know). I have stored procedures that run other procedures too; and I have offered a -c 500MB cash
which some have asked to decrease the cash to lower values for some cases (Which i don't either do know).
I will try to add page size to 4k, decrease the 500MB cash to 50MB (database size is 30MB) and trace it down again.
I wanted to post this in case you would still try to find out.
What do you think???

You have a corrupted database. No tweaking of switches is going to resolve that.
You need to do a full DBUNLOAD, DBINIT a fresh, empty .db file, and RELOAD the schema and data into that new database.

--
Paul Horan[TeamSybase]

John Vai

2005-04-12 21:35:58 UTC

Permalink

Thanks Paul for the reply; i will try to make the the things that you
suggested.

i will feed you back then

Thanks,

Post by Paul Horan[TeamSybase]

Post by John Vai
I got today a new assertion which is: 51901
Page for requested record not a table page or record not present in page.
I tried to search in the newsgroups about this specific assertion; what i
got is that they suggested to increase the page size to 4k.
I had some other posts from experts about problem causing the LONG
VARCHAR in tables. I have plenty of them in my table (why i don't know).
I have stored procedures that run other procedures too; and I have
offered a -c 500MB cash which some have asked to decrease the cash to
lower values for some cases (Which i don't either do know).
I will try to add page size to 4k, decrease the 500MB cash to 50MB
(database size is 30MB) and trace it down again.
I wanted to post this in case you would still try to find out.
What do you think???

You have a corrupted database. No tweaking of switches is going to resolve that.
You need to do a full DBUNLOAD, DBINIT a fresh, empty .db file, and RELOAD
the schema and data into that new database.
--
Paul Horan[TeamSybase]