Hi folks,
We're having a problem with Webspeed 2.1 and cgiip. Any help will be most welcome!
Configuration:
Local box(where I'm at): HP-UX Webspeed 2.1 with cgiip
Remote box: HP-UX Webspeed 2.1
We're using Webspeed 2.1 & cgiip thru Unix on the localbox to make queries on the remote unix box. The cgiip queries a PROGRESS dB on the remote box and returns with the output which is stored in a Buffer defined at runtime ($BUFFER).
The query passed directly from the localbox is something like this:
####################################################
export DLC=/u07/webspeed21
export WSPDLC=/u07/webspeed21
CONTENT_LENGTH=100000;export CONTENT_LENGTH
CONTENT_TYPE=application/x-www-form-urlencoded;export CONTENT_TYPE
REQUEST_METHOD=POST;export REQUEST_METHOD
WSPDLC=${WSPDLC:-/u07/webspeed21};export WSPDLC
export PATH=.:$DLC/bin:$PATH:$WSPDLC/bin
PROMSGS=$DLC/promsgs;export PROMSGS
PROPATH=$DLC;export PROPATH
PATH_INFO=/src_tree/dsi/pro/dsi.p;export PATH_INFO
SCRIPT_NAME=abcd;export SCRIPT_NAME
BUFFER='BUFFER0000=GETDATA1abcdata/main.pf custdb 22040323
0000011052690900 FRA &BUFFER0002=%3AEND%3A&'
echo "$BUFFER"|cgiip lmcx01 32320
####################################################
Problem: The above query works fine when run from local box for any user id, but fails totally for a particular production user id (mprod1). Now this is the default production id which is used by the production users and it has complete access to the directories, paths, files etc.... Also we've checked and double checked the parameters/config for this user id and compared with other id's and theres nothing different.
We're totally foxed by this one because the query works brilliantly from just about any other user id except mprod1.
We've totally checked out the DLC, PROPATH, PROMSGS etc and its all set correctly. One guess we have is that somehow this particular id (mprod1) has been marked out on the remote server and the broker does not service any requests made from this id.
On the remote server whenever a request is made by 'mprod1', the following happens:
WTB: Received a SIGCHLD from a child process. The dying process pid = 13874. (5743)
Requests from any other userid on the localbox is getting thru, but not mprod1!
Is it possible that the requests made by cgiip from the localbox are getting cached somewhere on the remote server with the login details etc?
How is webspeed able to distinguish the querying userid (it appears to be so!) and block its request?
How is it that only requests from the production id (mprod1) are getting blocked everytime?
I have been cracking my head on this one for a while now! Any input would be greatly appreciated!
Also it would be great if someone could explain how the cgiip object works out on Unix or points me to some usefule doc on Webspeed 2.1
Thanks in advance!
V Nair
We're having a problem with Webspeed 2.1 and cgiip. Any help will be most welcome!
Configuration:
Local box(where I'm at): HP-UX Webspeed 2.1 with cgiip
Remote box: HP-UX Webspeed 2.1
We're using Webspeed 2.1 & cgiip thru Unix on the localbox to make queries on the remote unix box. The cgiip queries a PROGRESS dB on the remote box and returns with the output which is stored in a Buffer defined at runtime ($BUFFER).
The query passed directly from the localbox is something like this:
####################################################
export DLC=/u07/webspeed21
export WSPDLC=/u07/webspeed21
CONTENT_LENGTH=100000;export CONTENT_LENGTH
CONTENT_TYPE=application/x-www-form-urlencoded;export CONTENT_TYPE
REQUEST_METHOD=POST;export REQUEST_METHOD
WSPDLC=${WSPDLC:-/u07/webspeed21};export WSPDLC
export PATH=.:$DLC/bin:$PATH:$WSPDLC/bin
PROMSGS=$DLC/promsgs;export PROMSGS
PROPATH=$DLC;export PROPATH
PATH_INFO=/src_tree/dsi/pro/dsi.p;export PATH_INFO
SCRIPT_NAME=abcd;export SCRIPT_NAME
BUFFER='BUFFER0000=GETDATA1abcdata/main.pf custdb 22040323
0000011052690900 FRA &BUFFER0002=%3AEND%3A&'
echo "$BUFFER"|cgiip lmcx01 32320
####################################################
Problem: The above query works fine when run from local box for any user id, but fails totally for a particular production user id (mprod1). Now this is the default production id which is used by the production users and it has complete access to the directories, paths, files etc.... Also we've checked and double checked the parameters/config for this user id and compared with other id's and theres nothing different.
We're totally foxed by this one because the query works brilliantly from just about any other user id except mprod1.
We've totally checked out the DLC, PROPATH, PROMSGS etc and its all set correctly. One guess we have is that somehow this particular id (mprod1) has been marked out on the remote server and the broker does not service any requests made from this id.
On the remote server whenever a request is made by 'mprod1', the following happens:
WTB: Received a SIGCHLD from a child process. The dying process pid = 13874. (5743)
Requests from any other userid on the localbox is getting thru, but not mprod1!
Is it possible that the requests made by cgiip from the localbox are getting cached somewhere on the remote server with the login details etc?
How is webspeed able to distinguish the querying userid (it appears to be so!) and block its request?
How is it that only requests from the production id (mprod1) are getting blocked everytime?
I have been cracking my head on this one for a while now! Any input would be greatly appreciated!
Also it would be great if someone could explain how the cgiip object works out on Unix or points me to some usefule doc on Webspeed 2.1
Thanks in advance!
V Nair
