IDX_SETS ERROR (AREV Specific)
At 24 NOV 1999 02:20:53PM Bill Wishart wrote:
Arron,
We met at the next to last roadshow presentation in Seatle.
I have gone through most of you suggestions for trouble shooting an IDX_SETS error that occurs during a select.
1. tried process again
2. tried FLUSH; GARBAGECOLLECT from TCL (this can't be done their runtime but i temporarily changed it)
3. tried the select from another workstation
4. Did'nt try it on a local copy
5. rebuilt indexes
Other things I've done
6. remaketable on both data and index based on veriflylh results
Current file stats are as follows:
Datafile
size lk 47,403,008 avg rows/group 4.73
ov 15,804,416 avg row len 166.60Framesize 1024
modulo 46292 load factor:
sizelock 0 .lk file .6393
free frames 0 .lk & ov .5907
Index file
size lk 9,211,904 avg rows/group 2.43
ov 5,779,456 avg row len 873.83Framesize 2048
modulo 4498 load factor:
sizelock 0 .lk file .7195
free frames 0 .lk & ov .6510
The frame size of the index file was at 1024 before the remake but I
still got the error.
What left?
frames of 64k of pointers.
Data file or index?
How do I find out?
What solutions are there?
How do I implement them?
At 30 NOV 1999 06:39PM Larry Wilson wrote:
What version of AREV are you using?
tardis_systems@yahoo.com
At 07 DEC 1999 03:41PM Bill Wishart wrote:
Version 3.12
At 07 DEC 1999 05:49PM akaplan@sprezzatura.com - [url=http://www.sprezzatura.com]Sprezzatura Group[/url] wrote:
What is the specific error? I've lost my cheat sheet so I can't match up specific errors to specific problems but generally LH structure isn't the problem.
Usually the problem happens when memory is running low and/or the select gets complicated. It's some sort of ASM buffer in the program getting overloaded.
The best solutions are to simplify the select into component parts, trying to find the one criteria the creates the smallest list.
After that, it's low level memory tweaking to try and maximize upper memory. I found a program the other day, TweakAll from http://www.abtons-shed.com which will give you some more control over memory during boot-up. Kernel toys will help some as well.
akaplan@sprezzatura.com
[/i]World leaders in all things RevSoft and archaic ARev stuff[/i]
At 16 DEC 1999 05:30PM Bill Wishart wrote:
—-FS445——————————————————————————————
IDX_SETS function 4 returned error code -30720
—-FS448—————————————————————————————–
Fatal internal error.]The select statement id not at all complex. It's a simple date range check "date ne "" and with date le "feb 1 1999"
Even if you reduce the range so that the number of keys selected is very small the error occurs.
At 18 DEC 1999 04:46PM akaplan@sprezzatura.com - [url=http://www.sprezzatura.com]Sprezzatura Group[/url] wrote:
These types of errors are pretty tough to track down. Best answers I have are searching through the message board for older threads and trying to see what you can do about maximising available memory before making the select.
akaplan@sprezzatura.com
At 10 JUL 2000 06:03PM Jocelyn Amon (finsol@ts.co.nz) wrote:
I am replying to this query at this late hour just in case anyone else is looking for an answer to this problem. The cause and solution to my variation on the problem does not apear to be covered by any of the responses so far.
I got the IDX_SETS error when a very helpful(!?) systems admin guy thought it might be a good idea to remove erase rights from the Revelation directories. The problem occurred when Revelation attempted to create and erase files from the sort directory.
So if you are having this problem, try deleting a file from the sort directory and see if you have rights to do so!
It took me a whole day to locate the problem as I was assured that nothing had been changed outside of Revelation!