Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.


Messages - Justin Derrick

Pages: 1 ... 108 109 110 111 112 [113] 114 115 116 117 118 ... 133
1681
You should be able to create a named query that produces the information you're looking for -- you can issue that query daily (by hand, or with "arsdoc query") and deliver it via (eMail?) from there.

From your description, this should satisfy most of your client's requirements.

-JD.

1682
iSeries / Re: Is Archive the correct term?
« on: September 14, 2012, 05:20:17 AM »
Hi Allen.

You'll likely want to store (aka 'archive') the file as AFPDS on the server, then convert it to PDF at the time it's retrieved by the user.  This is supported by 'transform' tools like IBM's APF2PDF.  I haven't checked with the latest versions of CMOD (which can split PDFs into resources and data, the same way CMOD does with AFP), but I still believe AFP is stored more efficiently (ie, it uses less space).

-JD.

1683
MP Server / Re: Copy TSM storage objects back to Cache ...
« on: September 08, 2012, 08:03:24 AM »
Backfilling a cache isn't easy.  You've already hit on a few of the points.

It's easier and faster to simply increase the cache retention for the AG that's underperforming, and reduce the caching for others.  This will improve performance over time, and requires practically no effort.

-JD.


1684
MP Server / Re: Recommended Inactivity Time Out Setting
« on: September 08, 2012, 07:53:45 AM »
As far as I know, the only impact an idle user has on server resources is RAM -- a few MB per user.  If the server is busy, this RAM would get swapped out to swap space.

Since CMOD isn't licensed on the 'concurrent user model' anymore, there's no advantage for licensing either.

I'd just set it to your company's standard inactivity timeout.

-JD.

1685
MP Server / Re: ars in oracle RAC
« on: September 06, 2012, 07:35:27 AM »
Siva...

Can you link to that article, or even upload it here?  Google can be fickle depending on where you are in the world...

Thx.

-JD.

1686
WEBI interface / Re: Load Balancing and Fail over in WEBi
« on: September 06, 2012, 07:34:15 AM »
Thanks for the excellent, point-form steps.  I'm sure this will be useful for our fellow users!

-JD.

1687
MP Server / Re: Sort Issue with 8.4 - 8.5
« on: August 31, 2012, 02:44:26 AM »
I didn't see any.  I think Greg mentioned it during his presentation at ODUG last year.

1688
MP Server / Re: Upgrade OnDemand 8.3 to 8.5
« on: August 31, 2012, 02:36:59 AM »
Hi Federico!

Thanks for posting this checklist -- but I have two questions...

Is step 14 for fixing the System Log bug in the 8.4 upgrade?

What is step 15 for?  Doesn't arsdb -uv do this for you?

-JD.

1689
MP Server / Re: ARS_NUM_DBSRVR - Efficient number of connections
« on: August 28, 2012, 05:33:15 PM »
What's the error message from DB2?  There are lots of situations that can create failed queries. Having too many database servers for CMOD is just going to gobble up some RAM that could be better used elsewhere.

When tweaking database parameters, I usually keep my changes to a maximum of 50% of the existing setting, and repeat this process until I find out if performance is improving.  If performance isn't improving, I roll back and look elsewhere.

1690
MP Server / Re: Turn off Login messge for a specific user ??
« on: August 28, 2012, 05:30:15 PM »
I'd issue an SQL delete command nightly that removes the previous day's logins...  Something like "where msg_num = 30 and userid like 'User%'".  You'd want to run re-orgs a little more frequently to clean up the swiss-cheese-effect you're creating in the database pages though.

1691
MP Server / Re: Sort Issue with 8.4 - 8.5
« on: August 28, 2012, 06:28:15 AM »
In your second sentence, you say QA is running 8.5, but lower down, you say it's running 8.4.0.3...  Which is which?

There have been sorting improvements in 8.5, so if your 8.5.0.5 server is working properly, but your 8.4.0.3 server is returning strange results, it's probably time to get that upgrade done!  ;)

-JD.

1692
I notified IBM of the problems yesterday -- they're working on it.

1693
It seems the documentation isn't quite on line yet.  I tried fetching some of it this morning, and all the PDF links end in IBM's default 404 page.

1694
Tips and Tricks / Re: Maximum number of rows per database table.
« on: August 21, 2012, 09:34:02 AM »
I always recommend that a table contains as many documents as are loaded within a month (10%-20% for growth), as long as that number is over 10 million rows.  I've recommended individual table sizes up to 150 million rows per table.  Conversations I've had with DB2 developers imply that this number could be much, much higher, but there are other issues that must be considered (backup and restore windows) before attempting it.

If you're in the position of loading millions of documents per month, segment expiration type is the recommended option, unless you're using things like Enhanced Retention Management or CFS.

1695
Tips and Tricks / Re: Calling a java program during indexing.
« on: August 21, 2012, 09:28:56 AM »
I can't imagine that this would perform terribly well.  What's the reason for calling Java?  What functionality are you trying to achieve that can't be done in C?

Pages: 1 ... 108 109 110 111 112 [113] 114 115 116 117 118 ... 133