Author Topic: Segmentation Fault with ARSDOC Get  (Read 781 times)

TRDFan

  • Guest
Segmentation Fault with ARSDOC Get
« on: October 18, 2021, 05:56:07 AM »
I'm working on an extract of just pulling 'raw' afp files from a very old, CMOD system that is being sunsetted by the end of the month. Obviously no support since this is V8.5.0.6

This particular load, was around 900MB, and contained 20 docs. When i tried retrieving the entire batch it was the same behavior. I figured i'd try retrieving just one or two docs. This makes me wonder, how were these being retrieved in the first place..? Once I start the extract.. /tmp fills up immediately. changing ARS_TMP doesnt fix it either, it still fills up /tmp.

Unfortunatly. There's pushback from our unix team to NOT increase /tmp..

Code: [Select]
bash-5.0$ arsdoc get -u admin -agcNv -G AG-i "where POLNO='12345' and rdate='18815'" -o test -h archive

10/18/21 08:46:14 : Starting arsdoc.  Version:  8.5.0.6

10/18/21 08:46:14 : arsdoc get -u admin -a -g -c -N -v -G AG -i where POLNO='12345' and rdate='18815' -o test -h archive

10/18/21 08:46:14 : Attempting login for userid 'admin' on server 'archive' ...

10/18/21 08:46:15 : Login successful

10/18/21 08:46:15 : Querying for application group 'AG' on server 'archive' ...

10/18/21 08:46:15 : Query successful

10/18/21 08:46:15 : Searching for documents in 'AG' ...

10/18/21 08:46:15 : Querying database with SQL string 'where POLNO='12345' and rdate='18815''

10/18/21 08:46:15 : Search successful

10/18/21 08:46:15 : 1 document(s) have been queried. Retrieving 1 document(s).

10/18/21 08:46:31 : (1): Retrieving document for userid 'admin' ...
Segmentation fault (core dumped)

TRDFan

  • Guest
Re: Segmentation Fault with ARSDOC Get
« Reply #1 on: October 19, 2021, 06:38:37 PM »
After some troubleshooting, and at the suggestion of another ODUG member, I executed my commands from another redhat server that I was days away from decommissioning, that had CMOD 9.5 installed on it. My extract process has been running flawlessly all afternoon, and significantly faster than my production extracts have been running.

Ironically, this unused box that probably has not been touched since 16/17', has probably 2-3 times the power as the current production CMOD system.