"JD" <MrJackDanielsBlack_at_Hotmail.com> wrote in
news:8n8Ne.9817$p%3.38423_at_typhoon.sonic.net:
>
> "Sybrand Bakker" <postbus_at_sybrandb.demon.nl> wrote in message
> news:h32ag1dl9ho44d7qvo0ta9fcmg53jfapii_at_4ax.com...
>> On Thu, 18 Aug 2005 21:50:20 GMT, "JD" <jd_at_hotmail.com> wrote:
>>
>>>We have a puzzling situation where doing a series of Inserts will
>>>block. The
>>>user has a transaction consisting of 4 inserts then has to confirm
>>>and we commit. If a second user starts the same sequence, the 4th
>>>insert will block
>>>until the first user commits. Our database people felt that there was
>>>a problem with foreign keys and indexes. Googling turns up lots of
>>>info on this.
>>>
>>>So, after an unsuccessful series of tests designed to locate the
>>>offending key, they disabled all of the foreign keys. Still blocked.
>>>So here's the question -- What other things will cause an insert to
>>>block? These are all new rows being inserted and each user has a
>>>different set of rows. And there
>>>are some triggers on a row insert.
>>>
>>>TIA,
>>>
>>>Mike P
>>>
>>
>> Did you actually look into v$lock by means of the Oracle provided
>> scripts?
>> Did you actually check out by means of the verious v$%wait% views
>> what your process is waiting for?
>>
>> --
>> Sybrand Bakker, Senior Oracle DBA
>
> Thanks for the quick response -- I'll have to run this by the database
> guys. Once their foreign key hypothesis vanished, they seem to lost
> interest. And they don't appear to have a firm grip on diagnosing the
> problem.
>
> Mike P
>
>
The blocking could be the result of having INITRANS too low.
This results in ENQUE waits
Received on Thu Aug 18 2005 - 21:11:19 CDT
