Hi,

I am stuck in a annoying issue called "illegal state Error 215" (attached screenshot).

I think problem started when a user killed a session (or did output to console), may be because of some locking issue.

Here the funny thing is, all user is able to login through BA mode but not through BW.

I read all the posts about Error 215 on BB as well as support site but no luck.

Does somebody faced similar issue or can put some light?

Thanks and Regards,

-Suhas

Hi Suhas,

in my tools manual there is no error 215 :-(
Sounds like a DB error - what do the logfiles of the database say?
Assuming that some tools table which is only needed when working with graphical interface (not the faintest idea which that might be) is messed up -
did you try to dump company 000? If that works without errors, I would bdbpost with (-k)ill option, thus dropping and recreating every table.

Sorry, no other idea comes to mind

good luck

Norbert

Thanks Norbert.

May be its DB error-

log.db2v5 shows -
2009-12-02[20:48:01]:E:bsp: ******* S T A R T of Error message *******
2009-12-02[20:48:01]:E:bsp: Log message called from /BAAN/view/port.6.1c.09/vobs/tt/servers/DB6000/db2_cli.c: #133 keyword: DB2 Error
2009-12-02[20:48:01]:E:bsp: Pid 385120 Uid 125 Euid 125 Gid 125 Egid 125
2009-12-02[20:48:01]:E:bsp: user_type S language 2 user_name bsp tty ote locale ISO88591/NULL
2009-12-02[20:48:01]:E:bsp: Errno 9 (Bad file number) bdb_errno 0
2009-12-02[20:48:01]:E:bsp:
2009-12-02[20:48:01]:E:bsp: dbs_errno = 0, Error 0 occurred:
2009-12-02[20:48:01]:E:bsp: Cannot allocate a DB2 Environment handle
2009-12-02[20:48:01]:E:bsp: Flushed at /BAAN/view/port.6.1c.09/vobs/tt/servers/DB6000/db2_cli.c : #134.
2009-12-02[20:48:01]:E:bsp: ******* E N D of Error message *******
2009-12-02[20:48:01]:E:bsp:
2009-12-02[20:48:01]:E:bsp: ******* S T A R T of Error message *******
2009-12-02[20:48:01]:E:bsp: Log message called from /BAAN/view/port.6.1c.09/vobs/tt/lib/dbs/dbs_server.c: #671 keyword: dbs_exec_XXX()
2009-12-02[20:48:01]:E:bsp: Pid 385120 Uid 125 Euid 125 Gid 125 Egid 125
2009-12-02[20:48:01]:E:bsp: user_type S language 2 user_name bsp tty ote locale ISO88591/NULL
2009-12-02[20:48:01]:E:bsp: Errno 0 bdb_errno 0
2009-12-02[20:48:01]:E:bsp: Log_mesg: Msg type = 0 returned error without setting dbs_errno
2009-12-02[20:48:01]:E:bsp: ******* E N D of Error message *******


Your droping and recreating table suggestion has risk, dont think we can adopt it.

Reboot

Dave

Already restarted Machine/DB/Application.

Have you tried to generate a dbs.log with:
-- -set DBSLOG=077777 in the field Command of the bw-client?
This should generate a (large) logfile called dbs.log at /home/bsp (that's the user if I interpreted your logfile correctly)

Have you altered $BSE/lib/ipc.info in any way?

No idea what "Cannot allocate a DB2 Environment handle" has to do with connecting with bw versus ba.

Please keep us informed and good luck to you

regards

Norbert

actually DBSLOG=1570 should do it.

http://www.davidmkerr.com/?p=140 for my howto on DBSLOG.

Dave

Thanks Norbert.

That option I already tried, attaching dbs_077.log.

And no, I havent toched to ipc_info.


Hi Dave

Thanks. It seem 07777 and 1570 works similar, attaching dbs.log file for more info.

Please provide more input.

For oracle some of the DBSLOG flags conflict each other so you don't always want to do 7777. I can't speak for other DBs.

However, in your case the driver fails before it can do anything.
>>db2_alloc_henv
>> - CLI function SQLAllocEnv returned value -1
FATAL /BAAN/view/port.6.1c.09/vobs/tt/servers/DB6000/db2_cli.c:#133

So something is up with your environment.

You said that the user can log in via ba but not bw, so that suggests that there is something in the unix environment that's messing him up.

When he logs in via BW is he using blogind? If he is, try rexec, if not try blogin.

Do you run any special bshell scripts that setup the environment before someone logs in?

Dave

Dave,

Yes, something is there at AIX level but I was wondering how can I trace/debug, unfortunately it didnt provided much info.

I tried both options rexec and baanlogin with & without blogind6.1 but no success.

No, we dont use any special bshell scripts.

Please provide other ways.

something is stopping DB2 CLI programs from starting. It's not a Baan issue at this point so no baan tracing will help.

You need to look to IBM for assistance.

Dave

Thanks Dave.

For your info. attaching actual log.bshell message, it showing some path does not exist.

2009-12-04[16:48:11]:E:yabiswal: ******* S T A R T of Error message *******
2009-12-04[16:48:11]:E:yabiswal: Log message called from /BAAN/view/port.6.1c.09/vobs/tt/mir/mir/bdb_fun.c: #1210 keyword: DB error
2009-12-04[16:48:11]:E:yabiswal: Pid 626822 Uid 217 Euid 217 Gid 125 Egid 125
2009-12-04[16:48:11]:E:yabiswal: user_type S language 2 user_name yabiswal tty ote locale ISO88591/NULL
2009-12-04[16:48:11]:E:yabiswal: Errno 0 bdb_errno 215 (Illegal state)
2009-12-04[16:48:11]:E:yabiswal: Log_mesg: Error 215 (Illegal state) on SELECT
2009-12-04[16:48:11]:E:yabiswal: ******* E N D of Error message *******
2009-12-04[16:49:52]:E:yabiswal:
2009-12-04[16:49:52]:E:yabiswal: ******* S T A R T of Error message *******
2009-12-04[16:49:52]:E:yabiswal: Log message called from /BAAN/view/port.6.1c.09/vobs/tt/mir/ds_link/li_init.c: #406 keyword: ds_close_server
2009-12-04[16:49:52]:E:yabiswal: Pid 626822 Uid 217 Euid 217 Gid 125 Egid 125
2009-12-04[16:49:52]:E:yabiswal: user_type S language 2 user_name yabiswal tty ote locale ISO88591/NULL
2009-12-04[16:49:52]:E:yabiswal: Errno 2 (A file or directory in the path name does not exist.) bdb_errno 0
2009-12-04[16:49:52]:E:yabiswal: Log_mesg: Detected Display Server Termination
2009-12-04[16:49:52]:E:yabiswal: ******* E N D of Error message *******

The error 215 is the real error. The 2nd error about path does not exist is the bshell reacting to the driver dieing.

Dave

Did u try latest set of BW / porting set (compatible to each other ) as per release notes. Is bw upgraded for all through central deployment and is not as per porting set requirement.

Hi Hitesh,

Thanks for Info.

Before three months only we have upgraded latest porting set 6.1c.09 as well as Infor's BECS as per release notes.

In fact, we worked almost three months on new PS and we suddenly encountered with these issues (may be locking issues).

Any suggestion?

Perhaps you could try removing the user defaults in session ttstpdeldeflt. I have found that sometimes the defaults cause errors that are very hard to track.

Best Regards,
Petteri

Petteri,

No..its not user default issue. Any other solution welcome.

Thanks

-Suhas

So you did use the new config for about three month?
But can't you remember what happened that day when the error first occured. Something must have changed.

Perhaps someone changed a config file on the LN / database server this day?
Was something else changed this day or a few days before?

I don't think it's a locking issue, cause normally when you restart the server all application and database locks are removed. One thing you can do if you believe it's a locking issue is search for files in /tmp , in $BSE_TMP or elsewhere that are changed around that time the error occured. There could be a lock file with some PID in it locking a needed process. But as I mentioned that's just a guess. I don't believe that this is the cause.

Yes, we used new config for three months without problem. As I said earlier, there could be locking issue but not sure.

AFAIK, nobody has changed config file since only our team has access.

You may right, it cant be locking issue but sorry to disagree restarting server cant clear locking always.

Thanks.

-Suhas