dvlahovic
12th June 2003, 07:30
Greetings to you all
I have a major problem for some time now and I'm running out of options how to solve it.
The thing is that in configuration BaanIVc4c Oracle817/Win I'm receiving error: Detected Display Server Termination.
I think it is the same thing as Francesco had once but what saw on that thread it didn't help me much to solve this.
It is not regular when it happens or how; it's like sometimes the whole session (Baan) dies but the connection to server (via RDP or VNC) is still working and sometimes, both connections are just gone.
I did check network interface, routing protocols, HW failures but everything seems to be working fine.
If try to ping server after it's diconnected I get no reply at all.
I also did some tracing but there was nothing usefull in bshell logs that I could be helped with.
Even if I work on server console the same thing happens.
So, now I need a piece of advice.
Did I forget about anything? What are the things that I should pay attention to?
Since Oracle is not my area of profession I would like to ask for patience with me.
I did attach a file with some info that it may be of some help (errlog, TNSconf ...).
Any help will be appreciated.
Regards,
dvlahovic
12th June 2003, 07:34
One thing that I forgot about is that server has last PS installed and also last SP (SP11).
BW is also latest.
The whole thing is running on Windows 2ks/SP3.
No IIS or addtional services.
Regards,
Francesco
18th June 2003, 23:47
Did you ever find that solution I mentioned to you?
The number is 105481 and it deals with settings on your windows TCP/IP that might be causing similar symptoms.
It also has links to other solutions about loss of connectivity.
dvlahovic
19th June 2003, 11:16
I saw the solution but it helped me only to solve the part of the problem. Connection over VPN is not reliable so during the transmission of packets ... some are lost and the VPN connection is broken.
The other part is that many sessions in Baan are still freezing during some operatons (printing, ...).
I still receive error: Database Server Termination.
Well, here comes the best part.
In event log there is also a error number ...232.
But I haven't found anything yet.
I don't know why but I keep thinking that it may be problem in Oracle. Maybe some parameters or just the fact that Oracle 8.1.7 will never work with Windows as it should.
Francesco
19th June 2003, 20:16
Error 232 is simply the "Connection Reset by Peer" error. I don't know why it is not included in any of the Baan error code tables I've ever seen.
I don't remember if I asked you this before, but does it happen in random sessions, or only in a limited group?
In your last reply, you mention a Database Server termination, when before you mentioned Display Server terminations. Was that a boo-boo or are you getting both errors (they're basically opposites sides of the spectrum)?
Another thing that doesn't quite rhyme is that you say you can't ping the server from a workstation after the disconnect occurs, but the disconnects occur on the server console as well.
Can you ping the server from the console after a disconnect?
Are you using latest versions of everything? BW client, porting set, database driver?
The 232 error would point at your BW, and the 'Display Server Termination' error kinda confirms that the disconnect is occurring from your client's end rather than the server.
This could also be a performance issue. If Oracle doesn't reply fast enough to a client's request, it will assume Oracle is dead, but that should give you the 'DAtabase Server Termination' error.
To make it really interesting (and I'm glad its you and not me, or I would substitute interesting with a word that may not be suitable for public display), you could have multiple issues going on at the same time.
Markus Schmitz
20th June 2003, 09:53
Hi Dario,
it sounds like you are in a big mess. Errors, which happen irregular are the most difficult to trouble shoot.
Still I think, you can approach this in a structured way.
First lets get your setup clear:
1. You have BaanIVc4 on Oracle on Windows2k running. All of this on a single server in host mode.
2. The users are also working on Win2k PC's working with the bwc GUI, not with Citrix or anything else.
3. Your Users are connected via local LAN.
Is this correct? Or are your users connected via VPN? If so, do you have Quality of Service from your provider?
Now to the actual error happening:
a) The error happens irregular and can not be reproduced.
b) You get the error "Detected Display Server termination"
Is this correct? Or do you get "Database server termination"?
Finally some analysis:
Both errors are actually completely different.
"Detected Display srever termination" means, that your bshell lost his connection to the GUI, either because the GUI crashed or because the network had a problem.
"Database server termination" means the bshell thinks the database server terminated, either because Oracle did not answer fast enough, or a network problem between Oracle and the bshell or Oracle crashed.
I assume for now, the first error happens. In this case, forget about your oracle settings, because they can not be the cause.
Additionally checking with VNC wether your Baan server can still be reached after the problem occured, might not tell you much, because the problem might be gone by then and because VNC might have different timeout, then the GUI.
What you have to do, is to measure proactively and continuosly, so that you have some data, when the problem occurs.
You can do this either with some network Mgt. Tool (open View and others.) You might not have such a tool. So you can try to make something up.
Write a small batch, which ping the server and writes the result to a log file. If you get "timeouts" in ping more or less, when you have a problem with the GUI, then you have just a network problem. You might have to install such a script in different parts of your network, as the error might be depending on your network topology.
Also you could look for a small tool, which can measure bandwidth over a period of time. I do not have one in mind, but I am sure, some are around.
All of the above assumes, a network problem (either HW or low level IP stack issues). But obviously you could have a problem with some software or configuration issues on the Baan server.
here you can forget in my opinion about your service pack of Baan. The only thing, that should matter is the Service Pack of Windows and the Protingset of Baan.
Also you might consider the following, if your setup permits. Forgive me, but this is not meant as a joke. As you are using windows, you solution might be to reboot your server every night. Stranger things happen in an windows environment. :-)
Regards
Markus
suhas-mahajan
20th June 2003, 10:21
Hi All,
This is already discussed earlier :
http://www.baanboard.com/baanboard/showthread.php?s=&threadid=9180
Same error, I am also getting, but rarely, earlier it was frequently. We have logged the case at support site, as per them, we have increased MAX_RETRIES, RETRY_TIME and BW_MAX_RETRIES < (digit) parameters. Problem not yet cured 100%. If anybody has some more solution, please share.
Thanks in advance.
-Suhas
dvlahovic
20th June 2003, 17:16
For Francesco:
what I receive is:
Env Baan (D:\Baan)
Prog bshell file \logic\mir\ds_link\li_init.c # 351
Keyword ds_close_server
Username baan type S language s
Process 0x3a98
Lasterror 0
Errno 0
bdberrno 0
Message Detected Display Server Termination
Env Baan (D:\Baan)
Prog bshell file \logic\mir\mir\main.c # 1501
Keyword signal handler
Username baan type S language s
Process 0xed4
Lasterror 997 (Overlapped I/O operation is in progress.)
Errno 0
bdberrno 0
Message Detected database server termination
Connection is broken randomly. It may happen while deleting a user or just want to exit session.
After the connection is broken I can still ping server and also there is no problem to log in.
And yes, I have installed all the latest "everything". PS, SP, BW ... it didn't help.
I had some thoughts on Oracle. Did check alert.log and there was error: TWO TASK SERVER, error = 12571.
But ... there are so many reason why it may happen. I really think that I did everything to reduce source of errors.
There is an otpion that Oracle is responding too slow. So, where can I define parameter to increase respond time?
-----------------------------------------------------------------------------------
For Markus Schmitz:
The configuration is all_in_one.
Application and DB on the same server.
Most of the users are working on w2k platforms but there are some with XP, or Win98.
Users are connectied via local LAN, 100Mbps.
Errors are happening without any logic or possible prediction.
If I want to compile a report; sometimes it work just fine and sometimes client just dies.
Strange thing is also that if I am workin on server console the same error occurs. So ... I know that if diconnect it from the NW, clien will crush. But I am sure that NW is fine be cause people from my company did measure the NW transsmisions and everything worked fine.
VPN connection is used only by me be casue I am the only one using remote access to server. I think that VPN is not the issue be cause as I have mentioned before, even if I work in local LAN or on the console of the server ... same thing happes.
Are there maybe any parameters to be set for and where for Oracle, BWC ...? My db_resource file looks like this:
dbsinit:01
nls_lang:slovenian_slovenia.ee8iso8859p2
nls_sort:binary
baan_oracle_prefetch:0
rds_full:2
ora_init:0111000
ora_max_array_fetch:30
ora_max_array_insert:20
lock_retry:0
ora_timeout:{999,999,999,999,999}
bdb_max_session_schedule:50
retained_cursors:60
max_free_cursors:64
I don't know if it sorrect but I don think that it is Ok.
Maybe your opinion.
-----------------------------------------------------------------------------------
For suhas-mahajan
I did check the link that you have provided me with. But there is nothing there that it was not done.
For the parameters: MAX_RETRIES, RETRY_TIME and BW_MAX_RETRIES ... it is true that I didn't enter them anywhere. It is be casue I don't know where to put them and be cause I don't know what values should I enter for each parameter.
As you can see it is all I have in db_resource file. So, are those parameters for client_side or server_side.
------------------------------------------------------------------------------------
Thank you all for trying to help me; I really appreciate that.
If ever this problem is solved, I invite you all to have one (or few) good Slovenian beers.
dvlahovic
20th June 2003, 17:28
Maybe one more detail that could help.
Configuration of BAAN server.
IBM x355
CPU: 2×2GHz PIV
RAM: 1,5 GB
HDD: C:\ RAID1 .... operating system
D:\ RAID1 .... Application
E:\ RAID5 .... DB
OS: Windows 2000 Server, SP3,
DB: Oracle 8.1.7
App: BaanIV c4, SP11
Porting Set: 6.1c.07
BWC: 6.1.3.70 (on the clients side)
Markus Schmitz
20th June 2003, 17:39
I think, one thing becomes obvious:
The error "Detected Display Server termination" is most likely a followup of the error "Database server terminated".
So I forget all thoughts about networking issues. And all thoughts about the bw client.
I would concentrate on the database and on the oracle driver configuration.
When the error in Baan happens, do you have corrsponding entries in the alertlog? Is the 12571 error always happening?
Second some comments to your db_ressource settings:
dbsinit:01
--> is ok, but we often use 021 nowadays
nls_lang:slovenian_slovenia.ee8iso8859p2
--> ok
nls_sort:binary
--> ok
baan_oracle_prefetch:0
--> never heard of this one, what is it doing?
--> remove it, if you do not know!
rds_full:2
--> ok
ora_init:0111000
--> ok
ora_max_array_fetch:30
--> why do you have this. We normally configure this parameter to be the same as rds_full!!!
--> try changing it.
ora_max_array_insert:20
--> dito
lock_retry:0
--> ok
ora_timeout:{999,999,999,999,999}
--> ok
bdb_max_session_schedule:50
--> Why do you have this?
retained_cursors:60
---> Should be a tuning issue only. Why 60?
max_free_cursors:64
---> dito
All guessing, of course :-)
Markus
dvlahovic
20th June 2003, 17:52
About error 12571:
log is written just sometimes.
Maybe in average, 2-3 per day.
It is not necessary for log to be written, even if error occurs.
About those parameters.
For all of them ... I am also not sure what they mean.
But as I was reading posts in baanboard I saw some comments that are describing about them.
So, just for test, I will erase the ones that you don't know what they are for and try with new db_resource configuration.
Thanks,
squaresh
21st June 2003, 02:37
I would try Runtime Tools, and shared memory restart
Sometimes, Tools is not setup correctly, and access to tools tables cause this kind of problem.
dvlahovic
30th June 2003, 17:28
Greetings
It seems like I did solve the problem.
To tell you the truth, there was not some specific action that has been helpful in solving the problem.
There were some parameters in db_reconfig and sqlnet.ora that has been changed, reconfigure the network connections,
Did RDD for tools....
It seems like all together did gave some effect.
There was nothing else I did to solve that. But is seems like it is working so ...
I would like to thank you all for you effort.
If one day in Slovenia .. please say so.
Beer is on me.
Cheers ... :)