cactid problems

Post by **raX** » Sat Apr 03, 2004 6:13 pm

mooseman wrote:By the way did you get my e-mail last week? I also cc'd my work e-mail but didn't receive it. So I was wondering if you got it.

Yep. I'll try and craft a reply sometime today.

-Ian

Morgan · Post by **Morgan** » Sat Apr 03, 2004 6:19 pm

raX wrote:I have a patch that might help fix the stuck RRDTool process problem. Just download the patch to your cactid directory and apply it like so:
Code: Select all
wget http://www.raxnet.net/downloads/cacti/patches/0.8.5a/cactid_rrdtool_pipe.patch
patch -p2 < cactid_rrdtool_pipe.patch
Make sure to recompile cactid after applying this patch and let me know how it goes.

-Ian

any chance u got a prepatched version somewhere . patches dont work right on solaris :/

Morgan

Post by **raX** » Sat Apr 03, 2004 6:28 pm

Morgan wrote:any chance u got a prepatched version somewhere . patches dont work right on solaris :/

Is there anything I can do when creating patches to ensure that they work on Solaris? Here is the link to the pre-patched version:

http://www.raxnet.net/downloads/cacti/p ... ched/rrd.c

-Ian

Morgan · Post by **Morgan** » Mon Apr 05, 2004 6:06 am

i'm really not sure how the patchign utility works under solaris, so i won't venture a guess.
Variosu people have posted output of the patching utility when we attempt applying a patch.
I think it might have something to do with the syntax the patch utility is expecting not matching what linux likes.

the joysd of solaris...
thanks for the .c though

i'll go compile this now.

dieter · Post by **dieter** » Tue Apr 06, 2004 2:05 am

I have been running cactid with the new .c file on a Win XP for 24 hours now and not a single stuck process. But let us not be too happy because this could be a lucky coincidence.

I will tell you next week if this 'hang' problem is really gone forever...

dieter · Post by **dieter** » Fri Apr 09, 2004 9:31 pm

ok, runs niiiiiiice now for a couple of days. I think we can say goodbye to the stuck processes forever (Win XP Prof, Cygwin 1.5.7)

Morgan · Post by **Morgan** » Sat Apr 10, 2004 3:28 am

solaris, still no stuck processes

Post by **raX** » Sat Apr 10, 2004 1:26 pm

Glad to hear that things seem to be working well now. I would be more than happy to put this bug behind us

.

-Ian

Morgan · Post by **Morgan** » Tue Apr 13, 2004 4:58 am

well unfortunately, it's not over yet.

:/

Code: Select all

host:/# ps -ef | grep cacti
    root 28385   191  0   Apr 10 ?        0:00 sh -c /usr/local/cactid/bin/cactid > /dev/null 2>&1
    root 17140   191  0   Apr 09 ?        0:00 sh -c /usr/local/cactid/bin/cactid > /dev/null 2>&1
    root 28386 28385  0   Apr 10 ?        0:01 /usr/local/cactid/bin/cactid
    root 17141 17140  0   Apr 09 ?        0:01 /usr/local/cactid/bin/cactid
host:/# ps -ef | grep rrd
    root 17143 17142  0   Apr 09 ?        0:00 /usr/local/bin/rrdtool -
    root 28387 28386  0   Apr 10 ?        0:00 sh -c /usr/local/bin/rrdtool -
    root 28388 28387  0   Apr 10 ?        0:00 /usr/local/bin/rrdtool -
    root 17142 17141  0   Apr 09 ?        0:00 sh -c /usr/local/bin/rrdtool -

dieter · Post by **dieter** » Wed Apr 14, 2004 3:57 pm

the same here

I didn't dare to mention it because I didn't want to spoil the fun (and I was not shure if this was due to cactid or due to a server reboot) but as now Morgan did post it, here is my post to say the problem isn't gone

bulek · Post by **bulek** » Thu Apr 15, 2004 4:15 am

Can you try my hint from another topic and report the results?
http://www.raxnet.net/board/viewtopic.p ... ght=#12034

- Piotr

Morgan · Post by **Morgan** » Thu Apr 15, 2004 4:40 am

well the bug isd evolving as it's patched Ian, first it was snmp queries getting stuck, then the scripts kept getting stuck.
now it seems to hang updating rrd's somewhere.

Post by **raX** » Thu Apr 15, 2004 11:24 pm

Morgan wrote:well the bug isd evolving as it's patched Ian, first it was snmp queries getting stuck, then the scripts kept getting stuck.
now it seems to hang updating rrd's somewhere.

I am glad to see that the bug is evolving, but there is definitely some greater issue that is causing these problems. One really strange thing is that these issues sort of crept up over night. I am pretty converative about changing core code in cactid, but it is still possible that something I changed is causing the problem. The problem also appears to be a lot more prevalent on Windows and Solaris platforms.

I find it interesting that the problem appears to go away when using a single thread. That doesn't surprise me really as problems like this tend to be thread related at the root. Perhaps the problem always existed but just wasn't noticed until cactid started running on Solaris and Windows. That is my best guess.

-Ian

Morgan · Post by **Morgan** » Fri Apr 16, 2004 3:45 am

Ian,
now that we've basically agreed that the bug is evolving, and doing so primarily on solaris and windows machine, the one thing i can see that all 3 variants of this bugs we've seen so far relates to rrdtool.
sure the CPU was getting hogged by different tasks each time, first by snmpget requests and then scripts in /cacti/scripts/ , but all of them have an rrdtool problem. i
just went back to the entire thread and each time we have a stuck rrdtool.

bug in rrdtool? incorrect argument passed somewhere when cactid tries to update rrd's? don't know, but it's a possibility that cactid isn't the problem here and that rrdtool is intead.

Post by **raX** » Sun Apr 18, 2004 11:06 pm

I wouldn't doubt if the bug was related to RRDTool somehow. Since I don't see myself building a Solaris test box anytime soon, I will try and setup my Windows machine with cactid. I have no doubt that I can debug this and possible find the source of the problem, I just need a reproducable test case.

-Ian

Cacti

cactid problems

Who is online