Any ideas on how I track down 3006 errors?

Archie Campbell

unread,

Apr 15, 2003, 12:48:53 AM4/15/03

to

Background:
Win2K server with P.SQL V8 (no SP1) Server Engine
24 workstations, mainly Win2K Pro, but some Win98, WinNT.
Application used MKDE (Btrieve stuff, not SQL). More specifically Visual
Dataflex 8.2 with ConnKit 1.3.
Network is 100Mbit ethernet.

Problem:
During the day various users get error 3006 (DFlex error#24006) and they
have to close and restart the application program.
It happens to different users. One user had it happen 5 times in a day.

3006: The MicroKernel router detected an invalid session.
. Connect has been closed
. abnormal network error in prior operation

Questions:
Any ideas how I start tracking this?
Are there tracing programs that would help?
Is there a log that I can look at?

Any thoughts would be appreciated.

Thanks
Archie Campbell

Cathy Delk

unread,

Apr 15, 2003, 10:06:21 AM4/15/03

to

First thing to do is get SP1.

If the problem persists and you haven't received another answer, contact
Pervasive's customer support. If the problem is on their side, the support
incident should be free.

"Archie Campbell" <arc...@accordsystems.com> wrote in message
news:AcMma.4344$945....@tor-nn1.netcom.ca...

Gordon

unread,

Apr 15, 2003, 10:40:34 AM4/15/03

to

Did you check the Pervasive log? You might find entries on the
serverlog stating invalid security tokens have been received. It's an
indication that you've got serious networking issues. In some cases
you may also see this in explorer, showing red crosses over the mapped
drives.

You could try using the autoreconnect function as a workaround. Be
sure to watch the timeout parameter though, as setting this too low
could in fact be the source of a statuscode 3006.

Gordon Bos
Q-RY Solutions
+31-(0)15-2564035

http://www.q-ry.nl/

Archie Campbell

unread,

Apr 15, 2003, 11:21:56 AM4/15/03

to

Gordon
Two very good thoughts.
Thank you, I will check it out.
Archie
"Gordon" <graf...@bigfoot.no.spam.please.com> wrote in message
news:qhacPgedGxPojV...@4ax.com...

Archie Campbell

unread,

Apr 15, 2003, 11:32:06 AM4/15/03

to

Cathy

I looked at the info on what was in SP1.
It seemed like not much, but I will take your advise and try it.
Archie

ps Sorry about the email.
I accidently clicked "Reply" instead of "Reply Group"
A

"Cathy Delk" <cd...@texsoft-consulting.com> wrote in message
news:xlUma.585$2_4.27...@newssvr11.news.prodigy.com...

Archie Campbell

unread,

Apr 15, 2003, 11:47:33 AM4/15/03

to

Oops
re "what was in SP1. It seemed like not much .."
Then I found
http://www.pervasive.com/support/updates/historyV8.txt

Archie

"Archie Campbell" <arc...@accordsystems.com> wrote in message

news:BDVma.4372$945....@tor-nn1.netcom.ca...

Archie Campbell

unread,

Apr 15, 2003, 12:47:29 PM4/15/03

to

Gordon
re: "You might find entries on the serverlog"
The serverlog?
I am aware of transaction logs and archival logs but no server logs.

Ahhh I have the "Archie is dumb" feeling, but I just had to ask ....

Thanks
Archie Campbell

"Gordon" <graf...@bigfoot.no.spam.please.com> wrote in message
news:qhacPgedGxPojV...@4ax.com...

Bill Bach

unread,

Apr 15, 2003, 2:40:19 PM4/15/03

to

Look at the PVSW.LOG file (in the Windows root directory, a.k.a. %WINDIR%) on
both the clients and server to see if anything is being reported in there...

3006 can sometimes be a secondary error, where the first error was not properly
trapped by the application.
Goldstar Software Inc.
Building on Btrieve(R) for the Future(SM)
Bill Bach
Bill...@goldstarsoftware.com
http://www.goldstarsoftware.com
*** Pervasive.SQL v2000i/v8 Service & Support Classes ***
2003: DC/IL: See our web site for details!

Gordon

unread,

Apr 15, 2003, 3:20:26 PM4/15/03

to

Sorry,

I meant the PVSW.LOG on the server (it's kind of a secret language,
you just skip what seems obvious......)

On Tue, 15 Apr 2003 10:47:29 -0600, "Archie Campbell"

Archie Campbell

unread,

Apr 16, 2003, 3:07:57 PM4/16/03

to

Bill and Gordon

Last night we did two things:
1. Installed SP1
2. Turned Auto Reconnect ON at the server and some workstations

This morning I read your postings and looked in pvsw.log
Its a brand new one with no entries prior to installing SP1 / reboot.
It has the following when the system was rebooted after the SP1 install
"Error initializing the SPX protocol. Error code: 6 NTDBSMGR.EXE"

Thanks for your thoughts

Archie

"Gordon" <graf...@bigfoot.no.spam.please.com> wrote in message

news:AlucPt4ofdwXdR...@4ax.com...

Archie Campbell

unread,

Apr 16, 2003, 5:40:27 PM4/16/03

to

Its a small 5k log, so I attached it.
(Not there. Maybe we reject attachments. Yeah, probably. So I will try
again, without the attachment)

pvsw.log looks like:
Date Time Component Process Process Name Computer Name
Type Category Msg ID Message
---------- -------- --------------- -------- --------------- ---------------
---- ---------- -------- --------------------------------------------------
-----------------------
04-15-2003 22:57:22 NTMKDE 000006E4 NTDBSMGR.EXE PACIFIC1
I Error initializing the SPX protocol. Error code: 6.
04-15-2003 22:57:32 NTMKDE 000006E4 NTDBSMGR.EXE PACIFIC1
W Resources allocated
04-15-2003 22:58:46 NTMKDE 000006E4 NTDBSMGR.EXE PACIFIC1
I Resources released
04-15-2003 23:01:25 NTMKDE 0000036C NTDBSMGR.EXE PACIFIC1
I Error initializing the SPX protocol. Error code: 6.
04-15-2003 23:21:55 NTMKDE 0000036C NTDBSMGR.EXE PACIFIC1
W Resources allocated
04-16-2003 12:29:14 NTMKDE 0000036C NTDBSMGR.EXE PACIFIC1
I Resources released
04-16-2003 12:29:22 NTMKDE 00000350 NTDBSMGR.EXE PACIFIC1
I Error initializing the SPX protocol. Error code: 6.
04-16-2003 12:29:22 NTMKDE 00000350 NTDBSMGR.EXE PACIFIC1
W Resources allocated
04-16-2003 12:37:37 W3COMSRV 00000350 NTDBSMGR.EXE PACIFIC1
I Client session timeout,
CID=ffffffff0000c0a801b807ee, SA=4644, TN=0000.
04-16-2003 12:38:22 W3COMSRV 00000350 NTDBSMGR.EXE PACIFIC1
I Client session timeout,
CID=ffffffff0000c0a801bd0546, SA=4644, TN=0000.
04-16-2003 12:39:52 W3COMSRV 00000350 NTDBSMGR.EXE PACIFIC1
I Client session timeout,
CID=ffffffff0000c0a80184064b, SA=4644, TN=0000.
<snip> deleted around 20 similar messages
04-16-2003 13:23:22 W3COMSRV 00000350 NTDBSMGR.EXE PACIFIC1
I Client session timeout,
CID=ffffffff0000c0a801b80413, SA=4644, TN=0000.

Last night we did the following:
. installed SP1
. turned on AutoReconnect at the server
This morning we turned on AutoReconnect at the clients
Life was rosy this morning

At 12:29 we accidently restarted all Pervasive services on the server.
Oops.
The users all reconnected.

Since then we are having lots of "Client session timeout" errors as shown in
the log.
Some users have had it happen two or three times.

Any thoughts would be appreciated.
Thanks

Archie

"Archie Campbell" <arc...@accordsystems.com> wrote in message
news:AcMma.4344$945....@tor-nn1.netcom.ca...

Archie Campbell

unread,

Apr 16, 2003, 6:12:07 PM4/16/03

to

On the server, in the "Monitor" utility, in the Communications section it
shows:

Connection Timeouts 40
Connection Recoveries 0

Is there something I may have set wrong that is causing it to never
"recover"?

Thanks

Archie

"Archie Campbell" <arc...@accordsystems.com> wrote in message

news:v6kna.4473$945....@tor-nn1.netcom.ca...

Bill Bach

unread,

Apr 17, 2003, 6:15:54 PM4/17/03

to

Did you enable PARC on BOTH the Client AND the server?

If the re-connect is not working, then it may just be that users are abnormally
terminating their workstations.

Goldstar Software Inc.
Building on Btrieve(R) for the Future(SM)
Bill Bach
Bill...@goldstarsoftware.com
http://www.goldstarsoftware.com
*** Pervasive.SQL v2000i/v8 Service & Support Classes ***
2003: DC/IL: See our web site for details!

Archie Campbell

unread,

Apr 17, 2003, 8:48:21 PM4/17/03

to

Bill
1. We have AutoReconnect active on both the client and the server.
We will reconfirm this at the client end.

2. Users are abnormally terminating ...
Nope. The local tech guy sat with them and observed.
A good thing to do.

He observed the following:
1. It only happens on the four NT workstations (all SP6a)
The Win2K machines are all OK.

2. It only happens when the workstation goes idle for 5 or more minutes
If the user keeps in the program and keeps working, no problem.
If they minimize the application, go do something else, when they come
back to the application it is disconnected

Thanks
Archie

"Bill Bach" <bb...@cncdsl.zerospam.com> wrote in message
news:3E9F279D...@cncdsl.zerospam.com...

Bill Bach

unread,

Apr 19, 2003, 8:36:17 AM4/19/03

to

Two possibilities:
1) Power management is somehow disabling the NIC card. Since NT doesn't have
reliable power management in the OS, you should verify that power management is
completely disabled in the BIOS configuration.
2) Are these machines using SPX to connect? If so, verify that SP6a is actually
installed. I had one customer tell me that SP6a was applied, but they actually
had the faulty SP6 installed. Run WINVER to verify this.

Archie Campbell

unread,

Apr 21, 2003, 5:55:13 AM4/21/03

to

Bill
I will check these.

Thanks very much for the ideas

Archie
"Bill Bach" <bb...@cncdsl.zerospam.com> wrote in message

news:3EA142C2...@cncdsl.zerospam.com...

Archie Campbell

unread,

Apr 21, 2003, 12:16:16 PM4/21/03

to

And now it happens on Win2000 Pro workstations too.
Not as frequently as WinNT, but ...
Hmmmm

Archie
"Archie Campbell" <arc...@accordsystems.com> wrote in message

news:FYHna.4533$945....@tor-nn1.netcom.ca...

okrasa

unread,

Apr 24, 2003, 3:44:10 AM4/24/03

to

I am pretty sure I received this error with Pervasive 7 when while my
program was running, some other application called reset on requester,
thus rendering my programs' connections closed.

--
Posted via http://dbforums.com

Archie Campbell

unread,

Apr 24, 2003, 9:09:53 PM4/24/03

to

Turning ON the client side caching seems to stop the disconnects.

Go figure.
Maybe when caching is on, the client periodically talks to the server, thus
maintaining the connection.
Kinda like a "keep busy" function.

Or, maybe not.

Anyway, the customer is happier now.

Thanks for everyone's suggestions.
Its much appreciated.

Archie

"Archie Campbell" <arc...@accordsystems.com> wrote in message
news:AcMma.4344$945....@tor-nn1.netcom.ca...