OpenBSD 3.7+CactiD hangs when running in crontab

[Trupaslu]

Hello,

At the moment i'm trying to get cactid to work. When i run cactid by hand:

Code: Select all

/usr/local/bin/php /data4/www/cacti/poller.php

I get:

Code: Select all

06/18/2005 08:15:27 PM - SYSTEM STATS: Time: 1.0289 s, Method: cactid, Processes: 1, Threads: 1, Hosts: 2, Hosts/Process: 2

Everything is fine and my graphs are getting updated. But if i try to run it in crontab the process hangs:

Code: Select all

root     17170  0.0  1.7  7128  8652 ??  S      8:20PM    0:00.29 /usr/local/bin/php /data4/www/cacti/poller.php
root      3868  0.0  0.1   388   304 ??  Is     8:20PM    0:00.00 /bin/sh -c /bin/sh
root      4773  0.0  0.1   188   672 ??  I      8:20PM    0:00.00 /usr/local/bin/rrdtool -

Hope someone can help me..

greet,

Sam

[TheWitness]

what version?

[Trupaslu]

I'm using version 0.8.6d

[TheWitness]

0.8.6e will be out very soon. RC23 is out now. It is located in the Announcements forum.

TheWitness

[Trupaslu]

For some reason it doesn't work.

Code: Select all

configure: error: cannot run /bin/sh config/config.sub

when i run /bin/sh config/config.sub:

Code: Select all

: not foundig.sub[5]:
: not foundig.sub[7]:
: not foundig.sub[26]:
: not foundig.sub[31]:
: not foundig.sub[39]:
: not foundig.sub[47]:
: not foundig.sub[54]:
: not foundig.sub[56]:
: not foundig.sub[69]:
: not foundig.sub[78]:
: not foundig.sub[81]:
config/config.sub[84]: syntax error: `in^M' unexpected

Thanks for your help..

Sam

[TheWitness]

In the RC's you need to run the following sequence of commands in order:

aclocal
auitoheader
autmake
autoconf
./configure
make
make install

Hope that helps.

TheWitness

[Trupaslu]

Yeah that gives the same error.

OpenBSD is my development server. Like i said earlier with cmd.php it works great!

The production server is a FreeBSD box. Cactid is working fine. What are your experiences with FreeBSD and cactid?

Currently we are running cacti-0.6.* and we are graphing about 7000 devices (seperated trees offcoure) but it is really slow because some trees hold about 600 graphs.

Thanks

Sam

[TheWitness]

In 0.8.6e, we are starting to Cache the tree's. Although, only as long as you set your browsers graphs to expire. The slowness is primarily due to the DHTML tree view stuff. We are looking for additional changes in the next release.

TheWitness

[TheWitness]

I don't have a clue as to why it's throwing that error.

TheWitness

[raX]

I got the same error on my Gentoo Linux machine when trying to run ./configure. I couldn't find the exact cause, but it works fine when building the official cactid package on my dev box.

If you want to confirm this under your environment, feel free to check out this build:

http://web-public.lan.raxnet.net/cactid ... .6e.tar.gz

-Ian

[brunom]

The solution I found to solve this problem is to run "dos2unix" on all files in the unzipped cactid source ... That seemed to fix the problem

/B

For some reason it doesn't work.

Code: Select all

configure: error: cannot run /bin/sh config/config.sub

when i run /bin/sh config/config.sub:

Code: Select all

: not foundig.sub[5]:
: not foundig.sub[7]:
: not foundig.sub[26]:
: not foundig.sub[31]:
: not foundig.sub[39]:
: not foundig.sub[47]:
: not foundig.sub[54]:
: not foundig.sub[56]:
: not foundig.sub[69]:
: not foundig.sub[78]:
: not foundig.sub[81]:
config/config.sub[84]: syntax error: `in^M' unexpected

Thanks for your help..

Sam

[brunom]

Hi,

I am having the same hanging processes when running cactid as the poller and invoked through crontab.

Code: Select all

root      2463  0.0  0.1  4280 1040 pts/2    S    17:57   0:00 less cacti.log
cacti     7459  0.0  0.0  2132  900 ?        S    18:25   0:00 /bin/bash -c php /var/www/cacti/poller.php > /dev/null 2>&1
cacti     7460  0.3  0.8 15568 8444 ?        S    18:25   0:00 php /var/www/cacti/poller.php
cacti     7466  0.0  0.0  2352  652 ?        S    18:25   0:00 /usr/local/bin/rrdtool -

My crontab looks like:

Code: Select all

# For Cacti
*/5 * * * * cacti php /var/www/cacti/poller.php > /dev/null 2>&1

I am running Cacti 0.8.6d and cactid rc23 (the latest latest one). When I run cactid from the command line it runs fine. When invoked from crontab, those processes mentionned above show up and I end up getting:

Code: Select all

06/20/2005 06:29:57 PM - POLLER: Poller[0] Maximum runtime of 296 seconds exceeded. Exiting.

I did an strace on the "poller.php" and this is what I see:

Code: Select all

[00a54004] lseek(4, 0, SEEK_CUR)        = 0
[00a54004] lseek(4, 0, SEEK_SET)        = 0
[00a54239] fcntl64(4, F_GETFL)          = 0 (flags O_RDONLY)
[00a534ae] fstat64(4, {st_dev=makedev(3, 2), st_ino=4194818, st_mode=S_IFREG|0644, st_nlink=1, st_uid=501, st_gid=0, st_blksize=4096, st_blocks=88, st_size=41963, st_atime=2005/06/20-18:10:00, st_mtime=2005/04/27-10:47:36, st_ctime=2005/06/20-17:40:30}) = 0
[00a5e3f3] mmap2(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0xb75e3000
[00a61aa5] _llseek(4, 0, [0], SEEK_CUR) = 0
[00a53e97] close(4)                     = 0
[00a5e4a1] munmap(0xb75e3000, 4096)     = 0
[00a54239] fcntl64(3, F_SETFL, O_RDWR|O_NONBLOCK) = 0
[00a53f0e] read(3, 0xa48fc08, 8192)     = -1 EAGAIN (Resource temporarily unavailable)
[00a54239] fcntl64(3, F_SETFL, O_RDWR)  = 0
[00a53f8e] write(3, "I\1\0\0\3select\n\t\tpoller_output.outp"..., 333) = 333
[00a53f0e] read(3, "\1\0\0\1", 4)       = 4
[00a53f0e] read(3, "\6", 1)             = 1
[00a53f0e] read(3, "\37\0\0\2", 4)      = 4
[00a53f0e] read(3, "\rpoller_output\6output\3\377\377\0\1\374\3\21\0\0", 31) = 31
[00a53f0e] read(3, "\35\0\0\3", 4)      = 4
[00a53f0e] read(3, "\rpoller_output\4time\3\23\0\0\1\f\3\3@\0", 29) = 29
[00a53f0e] read(3, "&\0\0\4", 4)        = 4
[00a53f0e] read(3, "\rpoller_output\rlocal_data_id\3\10\0\0"..., 38) = 38
[00a53f0e] read(3, "\37\0\0\5", 4)      = 4
[00a53f0e] read(3, "\vpoller_item\10rrd_path\3\377\0\0\1\375\3\1\0\0", 31) = 31
[00a53f0e] read(3, "\37\0\0\6", 4)      = 4
[00a53f0e] read(3, "\vpoller_item\10rrd_name\3\23\0\0\1\375\3\3@\0", 31) = 31
[00a53f0e] read(3, "\36\0\0\7", 4)      = 4
[00a53f0e] read(3, "\vpoller_item\7rrd_num\3\2\0\0\1\1\3!\0\0", 30) = 30
[00a53f0e] read(3, "\1\0\0\10", 4)      = 4
[00a53f0e] read(3, "\376", 1)           = 1
[00a53f0e] read(3, "\1\0\0\t", 4)       = 4
[00a53f0e] read(3, "\376", 1)           = 1
[00a186bd] time(NULL)                   = 1119255093
[009acf79] rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0
[009ace3e] rt_sigaction(SIGCHLD, NULL, {SIG_DFL}, 8) = 0
[009acf79] rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
[00a2e8bb] nanosleep({1, 0},  <unfinished ...>
Process 5712 detached

The stuff that caught my eye is the :

Code: Select all

[00a53f0e] read(3, 0xa48fc08, 8192)     = -1 EAGAIN (Resource temporarily unavailable)

I am not sure what the problem is ... it's strange ... Maybe when running cactid we need to invoke it from the command line instead the php poller?

[TheWitness]

Please try the release code. There are subtile changes with the handling of MySQL in the final release that were not there in rc23. Your code seems to be hanging around the MySQL calls.

TheWitness

[Trupaslu]

I'm using cacti-cactid-0.8.6d (Release)

Shall i wait till cacti-cactid-0.8.6e?

greet,

Sam

[TheWitness]

Should be out today.

Larry