Processing batch job as different user/different transaction time

[avillanueva]

I am just looking at areas to investigate. I have a batch job that
processes a list of WTParts. I was getting about 2 sec per job and I
was expecting faster rate. I used a different user account to speed
things up and noticed that the other user was less than 1 sec per job
and both were running in parallel. I do not think I was stressing out
the system but the question that remains is what would be different
about the users in the system (DB or LDAP) that could cause a difference
in processing times. Here is some info:

* Job involved Check in/out and attribute updates
* Parts were all in common library
* Noticed long times displaying updates table on Home - Overview
after processing (Most Recent view) - obviously, my account processed
lots of items
* My account is main context creator and we have 150 products +.
I'm into everything.
* System has not been restarted since jobs have run.



Some areas I think may be the cause:

* ACLs - Matrix for processing access (if I can check out or in)
is probably larger for my account than the other.
* DB - When my account is queried or used as a filter, more rows
are returned since it has more updates.



I can try some things on a test server but time is short and this is
more of a curiosity at the moment. Anyway, if you have any general
thoughts, love to hear them.

[RandyJones]

On 05/10/11 05:52, Villanueva, Antonio wrote:
>
> I am just looking at areas to investigate. I have a batch job that processes a
> list of WTParts.  I was getting about 2 sec per job and I was expecting faster
> rate.  I used a different user account to speed things up and noticed that the
> other user was less than 1 sec per job and both were running in parallel. I do
> not think I was stressing out the system but the question that remains is what
> would be different about the users in the system (DB or LDAP) that could cause
> a difference in processing times. Here is some info:
>
>     * Job involved Check in/out and attribute updates
>

Default check out preferences can make a difference - None vs Required vs All
If you have replica sites make sure the preferred file server preference is set
correctly.

>    *
>
>
>     * Parts were all in common library
>
>    *
>
>
>     * Noticed long times displaying updates table on Home -- Overview after
>       processing (Most Recent view) -- obviously, my account processed lots of
>       items
>

Browser cache settings can make a big difference. For example in FireFox the
browser.cache.check_doc_frequency option set at 3 (ootb) works faster than if
set to 1.

>    *
>
>
>     * My account is main context creator and we have 150 products +. I'm into
>       everything.
>     * System has not been restarted since jobs have run.
>
> Some areas I think may be the cause:
>
>     * ACLs -- Matrix for processing access (if I can check out or in) is
>       probably larger for my account than the other.
>     * DB -- When my account is queried or used as a filter, more rows are
>       returned since it has more updates.
>
> I can try some things on a test server but time is short and this is more of a
> curiosity at the moment. Anyway, if you have any general thoughts, love to
> hear them.
>
>
> ----------


--
------------------------------------------------------------------------
Randy Jones
Systems Administrator
Great Plains Mfg., Inc.
1525 E North St
PO Box 5060
Salina, KS USA 67401
email: -
Phone: 785-823-3276
   Fax: 785-667-2695
------------------------------------------------------------------------