bb-rep.sh & bb-snapshot.sh
list Bob Gordon
Hi all - I'm running into a problem creating the availability report and the snapshot on 4.3.0-beta2. Every time I run either command I get an "Internal Server Error". The apache logs show "Premature end of script headers" and the syslog shows that a buffer overflow was detected. The system its on is a Ubuntu 9.04 64b server. I have gone through the archives and based on what i did find i double-checked permissions(all were good).. Any ideas? The other option I have is to try to recreate the report manually.. Any ideas what datasets go into the report? thanks! -- --==[ Bob Gordon ]==--
list Asif Iqbal
▸
On Fri, Mar 12, 2010 at 5:16 AM, Bob Gordon <user-488dbf322a4e@xymon.invalid> wrote:
Hi all - I'm running into a problem creating the availability report and the snapshot on 4.3.0-beta2. Every time I run either command I get an "Internal Server Error". The apache logs show "Premature end of script headers" and the syslog shows that a buffer overflow was detected. The system its on is a Ubuntu 9.04 64b server.
I was having exact same error. I am running xymon version 4.4 beta. It is svn revision 6311. I compiled it on ubuntu 10.04. I see only google-chrome able to generate report. It is failing with FF 3.6.3 on ubuntu 10.04, FF 3.5.8 on opensolaris snv_134, and IE 8 on win xp I went ahead and changed the CFLAGS on build/Makefile.Linux like this before: CFLAGS = -g -O2 -Wall -Wno-unused -D_REENTRANT $(LFSDEF) $(OSDEF) after: CFLAGS = -g -O2 -Wall -D_REENTRANT $(LFSDEF) $(OSDEF) and recompiled using gcc version 4.4.3 and I don't see the "Premature end of script headers" any more. However, now I see a message in the browser, stays for a quick second, right before the report shows up. I can live with that pre-report message until I find a substitute for -Wno-unused. -Wno-unused does not show as a valid switch for gcc version 4.4.3
▸
I have gone through the archives and based on what i did find i double-checked permissions(all were good).. Any ideas? The other option I have is to try to recreate the report manually.. Any ideas what datasets go into the report? thanks! -- --==[ Bob Gordon ]==--
--
Asif Iqbal
PGP Key: 0xE62693C5 KeyServer: pgp.mit.edu
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?
list Rolf Biesbroek
After debugging together with some colleagues, I found a solution for this problem. In the source code of report.c and snapshot.c (in the map ./web) there is a declaration in the main function: "char htmldelim[20];". In the same main function we can find: " sprintf(htmldelim, "xymonrep-%u-%u", (int)getpid(), (unsigned int)getcurrenttime(NULL));" Because the size of "char htmldelim[20];" is not sufficient enough (requires around 25 or so) this process will crash. I changed this char value into 100 and after a "make && make install" command the issue seems to be resolved! Kind Regards, Rolf Biesbroek University Twente
▸
Hi all -
I'm running into a problem creating the availability report and the snapshot
on 4.3.0-beta2. Every time I run either command I get an "Internal
Server
Error". The apache logs show "Premature end of script headers" and the
syslog shows that a buffer overflow was detected.
The system its on is a Ubuntu 9.04 64b server.
I have gone through the archives and based on what i did find i
double-checked permissions(all were good)..
Any ideas?
The other option I have is to try to recreate the report manually..
Any
ideas what datasets go into the report?
thanks!
--
--==[ Bob Gordon ]==--
list Henrik Størner
▸
In <user-27452f8c0884@xymon.invalid> <user-47f84fd3870a@xymon.invalid> writes:
In the source code of report.c and snapshot.c (in the map ./web) there is a declaration in the main function: "char htmldelim[20];". In the same main function we can find: " sprintf(htmldelim, "xymonrep-%u-%u", (int)getpid(), (unsigned int)getcurrenttime(NULL));"
Because the size of "char htmldelim[20];" is not sufficient enough (requires around 25 or so) this process will crash.
Thank You! Very nice catch - good to have this done before the 4.3.0 release. There is also a potential crash that can be induced by feeding the report CGI a long "style" setting. I've fixed that as well in a commit just now. Regards, Henrik