Crashing Webservers CGI (OpenInsight Specific)
At 26 OCT 2000 12:36:58AM R Richardson wrote:
I have been having a problem with crashing
my webservers ability to process the OI CGI
program listed below. The webserver will still serve
pages it just won't run any oicgi.cgi. It will run
it for about a week and then start returning error
messages. This is on a pentium III with 128k and win98.
I am running my program with a link like this.
http://libertynatural.com/oicgi.exe/INET_CATLIST
It will crash the ability of both Website Pro and
the Sambar Server to process CGI programs.
Any help would be appreciated.
Roger
Here is my program.
function Inet_Catlist(Request)
$insert inet_equates
html=Content-type: text/html"
html = "
html =Customer Listing
"
OPEN ' CUSTOMERS' TO CUST ELSE PRINT 'cant open Web_Customers':STOP
OPEN 'DICT','CUSTOMERS' TO @DICT ELSE PRINT 'CANT OPEN DICT':STOP
SELECT CUST
LOOP:
READNEXT @ID ELSE GOTO FINREAD @RECORD FROM CUST, @ID ELSE GOTO FINhtml ={CUSTOMER}:""
GOTO LOOP
FIN:
html = "
return html
At 26 OCT 2000 04:07AM Oystein Reigem wrote:
R,
A couple of comments on that function. Probably nothing to do with your problem.
OPEN ' CUSTOMERS' TO CUST ELSE PRINT 'cant open Web_Customers':STOP
There's a leading space in the table name. But if it works one time it should always work.
But who do you want to tell if OPEN fails? If it's the web user you must return the error message as html, not print it on the server's printer. Perhaps it's this OPEN that for some reason fails after a while? Perhaps something happened to the table??
To return an error message to the user you can use the Inet_Msg function. E.g, do
OPEN ' CUSTOMERS' TO CUST ELSE
Message=cant open Web_Customers'
return Inet_Msg( Request, Message,
,
)END
OPEN 'DICT','CUSTOMERS' TO @DICT ELSE PRINT 'CANT OPEN DICT':STOP
Same comment about error handling.
- Oystein -
At 26 OCT 2000 04:14AM Tim Marler @ Prosolve Software (UK) wrote:
Try setting the Inet_Gateway to Local Development Mode. When you get an error on the server, OI should break to the debugger and you'll be able to see where the problem is being caused.
Tim
tmarler@prosolve.co.uk @ Prosolve Software
At 26 OCT 2000 04:20AM Oystein Reigem wrote:
R,
It's much easier with problems that occur consistently, and from day 1. To get a handle on your problem perhaps you could log what's happening when your function runs. My assumption is now that your function really does start to run, but then crashes or hangs.
One way you can log is to have your inet function display messages in the Inet_Gateway window. See .
- Oystein -
At 26 OCT 2000 04:30AM Oystein Reigem wrote:
Tim,
Thanks for the tip from me too.
But is this good advice in all error situations? What if there's a bug in the programming - a minor one, and one that's seldom encountered, but it causes a break to the debugger. Then all execution would stop, and all users would be affected - not just one. Right? Just so I understand what it's good for and what not.
- Oystein -
At 26 OCT 2000 08:43AM [url=http://www.sprezzatura.com]The Sprezzatura Group[/url] wrote:
While the STOP command might be the problem, seeing as how it's an invalid opcode in OpenInsight and does not compile. The print statement might cause a few issues as well.
Our past working with OICGI found that it occasionally stops responding. When this happens, you can try bringing the INET_GATEWAY to the foreground, as that has been known to restart access.
If that does not work, then rebooting the server is requied.
World Leaders in all things RevSoft
At 26 OCT 2000 10:48AM Donald Bakke wrote:
Roger,
I don't know if this is the same problem or not, but we used to have a problem with our OICGI apps no longer responding after about 300 hits. This was very consistant. We finally reported it to Revelation and WinWin identified the problem with a similar one they had already resolved.
For whatever reason, Win98 OEM versions cause this problem. Check the Properties of your My Computer and see if OEM is in the registration code. If so, then this might be your problem. We switched to NT 4.0 and have not have a problem ever since.
dbakke@srpcs.com
At 26 OCT 2000 11:50AM Oystein Reigem wrote:
At 26 OCT 2000 11:53AM Oystein Reigem wrote:
saw that STOP too but forgot to comment on it. but it says colon STOP not semicolon STOP. so it can't be the STOP command. must be a variable R forgot to set?
- o -
At 27 OCT 2000 05:30AM Tim Marler @ Prosolve Software (UK) wrote:
Oystein,
I use it only when developing. By the time your app goes live there won't be any bugs in the code anyway
Seriously, you will need to trap errors and handle them accordingly. I always run live apps in Runtime mode.
Tim
tmarler@prosolve.co.uk @ Prosolve Software
At 27 OCT 2000 07:33AM Oystein Reigem wrote:
Tim,
By the time your app goes live there won't be any bugs in the code anyway
I blinked while reading and suddenly thought it said . How appropriate.
![]()
- Oystein -
At 27 OCT 2000 08:46AM Oystein Reigem wrote:
Tim,
Btw - do you think bwg could be Welsh for [/b]bug[/b]?
Also - what do you think of the Debugger I just got up on my screen? Don't think I ever saw it in that mood before. Must have felt threatened by resources being low, and decided to put up as good a defence as it managed. Got out all its bug killing tools…
- Oystein -