E_dm9c75_dm2d_close_tcbbusy

[omaro_ing]

Hi,

we have migrated our DBMS server Ingres2006 to Ingres10 with OS RedHat5.5 , the performance was enhanced, but i have an error that repeated often :


LOCALHOST ::[36677 , 4158 , 00002aaad66fa200, dm2d.c:2997 ]: Thu Feb 9 17:26:47 2012 E_DM9C75_DM2D_CLOSE_TCBBUSY Table Descriptor unexpectedly busy at database close time. The TCB for table (detail_fac_bt_pay2007, sgc) of database sgcbt was found referenced while the database was being closed: tcb_ref_count 1, tcb_valid_count 1, tcb_status 0x00000001.
LOCALHOST ::[36677 , 4158 , 00002aaad66fa200, dm2d.c:2997 ]: Thu Feb 9 17:26:47 2012 E_DM9C75_DM2D_CLOSE_TCBBUSY Table Descriptor unexpectedly busy at database close time. The TCB for table (detail_fac_bt, sgc) of database sgcbt was found referenced while the database was being closed: tcb_ref_count 1, tcb_valid_count 1, tcb_status 0x00000001.
LOCALHOST ::[36677 , 4158 , 00002aaad66fa200, dm2d.c:3052 ]: Thu Feb 9 17:26:47 2012 E_DM9270_RELEASE_TCB Error occurred releasing a TCB.
LOCALHOST ::[36677 , 4158 , 00002aaad66fa200, dm2d.c:2795 ]: Thu Feb 9 17:26:47 2012 E_DM9267_CLOSE_DB Error occurred closing a database.
LOCALHOST ::[36677 , 4158 , 00002aaad66fa200, scddbfcn.c:387 ]: Thu Feb 9 17:26:47 2012 E_SC010E_DB_DELETE Error deleting database. Name: sgcbt Owner: sgc Added Id D022E040
LOCALHOST ::[36677 , 4158 , 00002aaad66fa200, dm2d.c:3129 ]: Thu Feb 9 17:26:47 2012 E_DM003F_DB_OPEN Database is open.
LOCALHOST ::[36677 , 4158 , 00002aaad66fa200, dmcclose.c:406 ]: Thu Feb 9 17:26:47 2012 E_DM0087_ERROR_CLOSING_DB Error closing database in server.
LOCALHOST ::[36677 , 4158 , 00002aaad66fa200, scsdbfcn.c:1087 ]: Thu Feb 9 17:26:47 2012 E_SC0122_DB_CLOSE Error closing database. Name: sgcbt Owner: sgc
LOCALHOST ::[36677 , 4158 , 00002aaad66fa200, scsdbfcn.c:1094 ]: Thu Feb 9 17:26:47 2012 E_SC010D_DB_LOCATION Database Location Name: $default Physical Specification: /bdd/Ingres10/IngresA4/ingres/data/default/sgcbt Flags: 00000003
LOCALHOST ::[36677 , 4158 , 00002aaad66fa200, scsinit.c:4721 ]: Thu Feb 9 17:26:47 2012 E_DM0087_ERROR_CLOSING_DB Error closing database in server.

OCALHOST ::[36677 , 4158 , 00002aaad66fa200, scsdbfcn.c:859 ]: Sat Feb 11 02:00:02 2012 E_SC0121_DB_OPEN Error opening database. Name: sgcbt Owner: sgc Access Mode: 00000002 Flags 40000000
LOCALHOST ::[36677 , 4158 , 00002aaad66fa200, scsdbfcn.c:868 ]: Sat Feb 11 02:00:02 2012 E_SC010D_DB_LOCATION Database Location Name: $default Physical Specification: /bdd/Ingres10/IngresA4/ingres/data/default/sgcbt Flags: 00000003
LOCALHOST ::[36677 , 4158 , 00002aaad66fa200, scsinit.c:7935 ]: Sat Feb 11 02:00:02 2012 E_DM0133_ERROR_ACCESSING_DB Error accessing database due to invalid internal information.
Access this
database from another server or shut this server down and start a new one.


what's a matter?

Thanks.

[kschendel]

A table control block is in a state that normally indicates "being used by a user session" at database close time. Since the database is being closed, this should not happen. I guess this is an internal bug, although I don't believe I've heard of it being reported elsewhere.

I think you will need tech support help for this one. As a temporary workaround, you can try avoiding the database close -- meaning, keep a session open accessing the database. The session doesn't need to do anything other than connect, you don't have to issue any queries.

[geraintjones]

Hi,
I think we've seen this sequence of errors before.
In our case the session which prevented the database being closed was running "copy table...into..." over Ingres/Net from a remote host, and had become stuck in a CNDWAIT state.
Unfortunately, we weren't able to reproduce the problem in a test system afterwards.
Do you have any details about the connected sessions in your case - eg a "format all" output from iimonitor?
Geraint

[jruffer]

We had a similar problem which occurred when the remote host was 9.2. This was fixed by patch 14128 for 9.2.1.

Jeremy

[geraintjones]

Jeremy, that's interesting. Just to confirm - the problem you saw was with a 9.2.0 client connecting to a 10.0 server? That fits in with the environment in which our problem occurred.
Also, were you able to reliably reproduce the problem? As we haven't been able to, it could be tricky for us to verify that upgrading the client to a patched 9.2.1 fixes it.
TIA ,Geraint

[jruffer]

Yes, it was a 9.2 client connecting to a 10.0 server. The problem also occurred with a 9.2.1 client.

It didn't happen everytime for me. I ended up finding a table that it did happen with and sent that as a testcase. I might be able to send it to you but would have to spend time making sure that I wasn't sending anything confidential.

There was a workround which was to set vch_compression = off in the server.

Jeremy

[geraintjones]

Originally Posted by jruffer

There was a workround which was to set vch_compression = off in the server.

Hmmm, this makes me doubt that we've seen exactly the same problem. In our case II_VCH_COMPRESS_ON=N had been set (but on the client side).

Jeremy, I really wouldn't ask you to hack around with the data in one of your test cases on our account. We haven't seen this problem for the last 6 or 7 months anyway and the server side has been patched several times in the meantime.

Not sure whether any of this helps the OP in any case - apologies if your thread got a bit hijacked there!

Geraint

[omaro_ing]

Hi,
inside of client we have openroad 3.5, but i observe that this problems happens in outside hours of working like night or in morning early and server ingres10 dont respond after this problem we must to restart it

thank you for your help.