From: Dan Holmes on
On 3/18/2010 12:45 PM, Kalen Delaney wrote:
> You still haven't told me what version you are using, or what query you
> are running to get the numbers you are showing. Although the columns
> names are a big clue, it's more polite to not make the people helping
> you try to guess. It still looks like the big number in
> resource_associated_entity_id is the partition_id. I think you didn't
> show all the rows. There should be some for OBJECT locks, and then the
> resource_associated_entity_id would be the object_id.
>
i didn't show all the rows there are 1770 of them.

> There is no way to hijack an open transaction, but you can kill it, and
> that will rollback the transaction and release the locks.
>

I am sorry. This is happening live and i am trying to gather how to fix this as well as manage the problem.,

this produced the output
Select *
FROM sys.dm_tran_locks
WHERE request_Session_ID = 187


your guess was right about the value

Select object_name(p.object_id), *
FROM sys.dm_tran_locks tl
LEFT JOIN sys.partitions p ON tl.resource_associated_entity_id = p.Partition_id
WHERE request_Session_ID = 187
and resource_database_id = 42


showed me what tables were affected.

i just tested my second theory

BEGIN TRAN;
UPDATE tblAgency SET RMqsserverportnumber = 12345
select * from sys.dm_tran_session_transactions

--stop here
--now i modified the spAddressSelect proc below
--and stuck in an extra commit.
Declare @ContextStr varchar(255)
set @ContextStr = dbo.syscontextstr()
exec [spAddressSelect]
@ContextStr ,
3,
1
select * from sys.dm_tran_session_transactions
--after running the transaction was gone and committed.

SELECT * FROM tblAgency
From: Uri Dimant on
Dan
Adam Machanic has written a great stored procedure to track down such
problems, please visit on www.sqlblog.com and look for
who_is_active_v8_82.sql



"Dan Holmes" <dan.holmes(a)routematch.com> wrote in message
news:uumZ3urxKHA.1548(a)TK2MSFTNGP02.phx.gbl...
> On 3/18/2010 12:45 PM, Kalen Delaney wrote:
>> You still haven't told me what version you are using, or what query you
>> are running to get the numbers you are showing. Although the columns
>> names are a big clue, it's more polite to not make the people helping
>> you try to guess. It still looks like the big number in
>> resource_associated_entity_id is the partition_id. I think you didn't
>> show all the rows. There should be some for OBJECT locks, and then the
>> resource_associated_entity_id would be the object_id.
>>
> i didn't show all the rows there are 1770 of them.
>
>> There is no way to hijack an open transaction, but you can kill it, and
>> that will rollback the transaction and release the locks.
>>
>
> I am sorry. This is happening live and i am trying to gather how to fix
> this as well as manage the problem.,
>
> this produced the output
> Select *
> FROM sys.dm_tran_locks
> WHERE request_Session_ID = 187
>
>
> your guess was right about the value
>
> Select object_name(p.object_id), *
> FROM sys.dm_tran_locks tl
> LEFT JOIN sys.partitions p ON tl.resource_associated_entity_id =
> p.Partition_id
> WHERE request_Session_ID = 187
> and resource_database_id = 42
>
>
> showed me what tables were affected.
>
> i just tested my second theory
>
> BEGIN TRAN;
> UPDATE tblAgency SET RMqsserverportnumber = 12345
> select * from sys.dm_tran_session_transactions
>
> --stop here
> --now i modified the spAddressSelect proc below
> --and stuck in an extra commit.
> Declare @ContextStr varchar(255)
> set @ContextStr = dbo.syscontextstr()
> exec [spAddressSelect]
> @ContextStr ,
> 3,
> 1
> select * from sys.dm_tran_session_transactions
> --after running the transaction was gone and committed.
>
> SELECT * FROM tblAgency