rrdfiles not getting updated in 0.8.6f

[Burvil]

Hi,

I have a clean install of cacti 0.8.6f, fully patched, i.e. with all patches at http://www.cacti.net/download_patches.php. I did a clean install after trying to upgrade from 0.8.5a, which seemed to be OK, but wasn't updating the rrdfiles (same problem as now, with clean install). Anybody have any thoughts on why the rrdfiles aren't being updated?

The poller (cmd.php) is running, seems to be collecting data. This is from one round of data collection in cacti.log (set to highest level, debug):

Code: Select all

08/15/2005 05:10:03 PM - CMDPHP: Poller[1] WARNING: PHP version 4.3 or above is recommended for performance considerations.
08/15/2005 05:10:03 PM - CMDPHP: Poller[0] Host[1] SNMP: Device does not require SNMP
08/15/2005 05:10:03 PM - CMDPHP: Poller[0] Host[1] No host availability check possible for '127.0.0.1'.
08/15/2005 05:10:04 PM - CMDPHP: Poller[0] Host[1] DS[3] CMD: perl /var/www/html/cacti/scripts/linux_memory.pl MemFree:, output: 62744
08/15/2005 05:10:04 PM - CMDPHP: Poller[0] Host[1] DS[4] CMD: perl /var/www/html/cacti/scripts/linux_memory.pl SwapFree:, output: 2067568
08/15/2005 05:10:04 PM - CMDPHP: Poller[0] Host[1] DS[5] CMD: perl /var/www/html/cacti/scripts/loadavg_multi.pl, output: 1min:0.67 5min:0.50 10min:0.56
08/15/2005 05:10:04 PM - CMDPHP: Poller[0] Host[1] DS[6] CMD: perl /var/www/html/cacti/scripts/un http://www.cacti.net/download_patches.php,ix_users.pl, output: 2
08/15/2005 05:10:05 PM - CMDPHP: Poller[0] Host[1] DS[7] CMD: perl /var/www/html/cacti/scripts/unix_processes.pl, output: 135
08/15/2005 05:10:08 PM - CMDPHP: Poller[0] Host[2] SNMP: Host responded to SNMP
08/15/2005 05:10:08 PM - CMDPHP: Poller[0] Host[2] RECACHE: Processing 2 items in the auto reindex cache for 'fileserver'.
08/15/2005 05:10:13 PM - CMDPHP: Poller[0] Host[2] DS[14] SNMP: v1: fileserver, dsname: cpuCPInterrupts, oid: .1.3.6.1.4.1.789.1.2.1.11.0, output: 1
08/15/2005 05:10:14 PM - CMDPHP: Poller[0] Host[2] DS[15] SNMP: v1: fileserver, dsname: sysUpTime, oid: .1.3.6.1.2.1.1.3.0, output: 302025289
08/15/2005 05:10:16 PM - CMDPHP: Poller[0] Host[3] DS[9] SNMP: v1: fileserver, dsname: netappCacheAge, oid: .1.3.6.1.4.1.789.1.2.2.23.0, output: 154
08/15/2005 05:10:17 PM - CMDPHP: Poller[0] Host[3] DS[10] SNMP: v1: fileserver, dsname: netappCIFSReads, oid: .1.3.6.1.4.1.789.1.7.3.1.1.5.0, output: 61959990
08/15/2005 05:10:18 PM - CMDPHP: Poller[0] Host[3] DS[11] SNMP: v1: fileserver, dsname: netappCIFSReads, oid: .1.3.6.1.4.1.789.1.7.3.1.1.6.0, output: 22592664
08/15/2005 05:10:19 PM - CMDPHP: Poller[0] Host[3] DS[12] SNMP: v1: fileserver, dsname: netappCPUBusy, oid: .1.3.6.1.4.1.789.1.2.1.3.0, output: 10
08/15/2005 05:10:21 PM - CMDPHP: Poller[0] Host[3] DS[13] SNMP: v1: fileserver, dsname: netappNFSOps, oid: .1.3.6.1.4.1.789.1.3.2.2.1.0, output: 450040604
08/15/2005 05:10:21 PM - CMDPHP: Poller[0] Time: 17.9125 s, Theads: N/A, Hosts: 3

So, I know the systems are in fact responding to the snmp requests, but that
the rrd files are not getting updated. Instead of 4:55PM or later, it shows the time the rrdfile was created.

Code: Select all

[root@cactiserver rra]# pwd
/var/www/html/cacti/rra
[root@cactiserver rra]# ls -lt | head
total 1673
-rw-r--r--    1 cactiuse cactiuse    95324 Aug 15 16:03 fileserver_template1_sysuptime_15.rrd
-rw-r--r--    1 cactiuse cactiuse    95324 Aug 15 16:02 fileserver_template2_netappcpubusy_12.rrd
-rw-r--r--    1 cactiuse cactiuse    95324 Aug 15 16:02 fileserver_template2_netappcacheage_9.rrd
-rw-r--r--    1 cactiuse cactiuse    95324 Aug 15 16:02 fileserver_template2_netappnfsops_13.rrd
-rw-r--r--    1 cactiuse cactiuse    95324 Aug 15 16:01 fileserver_template2_netappcifsreads_11.rrd
-rw-r--r--    1 cactiuse cactiuse    95324 Aug 15 16:01 fileserver_template2_netappcifsreads_10.rrd
-rw-r--r--    1 cactiuse cactiuse    95324 Aug 15 16:01 fileserver_template1_cpucpinterrupts_14.rrd
-rw-r--r--    1 cactiuse cactiuse   375572 Aug 15 15:48 fileserver_template1_kbytestotal_8.rrd
-rw-r--r--    1 cactiuse cactiuse    95324 Aug 15 12:54 fileserver_fileserver_disk_read_28.rrd
[root@cactiserver rra]#

I did notice something earlier where I had to manually create the datasource by turning on debug mode in the datasources and running rrdtool create from the command line, cutting and pasting the command from browser to the terminal. :

Looked at http://www.cacti.net/download_patches.php, following instructions and installed all patches:

Code: Select all

patch -p1 -N < html_mouseover_title.patch
patch -p1 -N < multi_field_output_script_server.patch
patch -p1 -N < old_net_snmp_command_options.patch
patch -p1 -N < poller_invalid_data_sources.patch
patch -p1 -N < script_server_buffer_size.patch
patch -p1 -N < unix_syslog_compatibility.patch
patch -p1 -N < graph_view_preview_mode_filter.patch
patch -p1 -N < poller_reindex_cache_orphans.patch
patch -p1 -N < safari_zoom.patch
patch -p1 -N < ldap_copy_user_problem.patch
patch -p1 -N < script_server_file_header.patch

I am using rrdtool 1.0.33. I poked around in the forums, and found the link http://forums.cacti.net/viewtopic.php?t=8583, but I don't think that's necessarily it, since this happened right after I upgraded.

[BSOD2600]

1) If you want to stick with the old rrdtool 1.0.x, thats fine. I suggest you use the latest version though.

2) You are running poller.php, right?

3) What versions of software are you running?

4) Was the rrd files initially created under a different username that the cron job is running now?

[Fred]

I am using rrdtool 1.0.33. I poked around in the forums, and found the link http://forums.cacti.net/viewtopic.php?t=8583, but I don't think that's necessarily it, since this happened right after I upgraded.

The problem you're describing is exactly the same as what's in that thread. Or this one has a bit more information.

To solve this on my system, I followed the Witness's instructions to comment out lines in poller.php that contain "rrd_init" or "rrd_close" (3 lines total, iirc). Been running like this for about a month now with no noticeable ill effects.

[Burvil]

Cool, thanks. Yeah, I changed my cron job to poller.php, and it's updating now. I guess it must have changed from cmd.php in a version change? Odd, in the GUI (console -> Configuration -> Settings -> Poller) it still only shows cmd.php and cactid as the pollers. I guess this is a bug?

[BSOD2600]

No, not a bug. A design change.

Now cacti checks devices to see if they're up or down. That way it won't waste time trying to poll down boxes and waste time. Poller.php will invoke either cmd.php or cactid.exe, which ever you are using.

[cacti_fod]

A design change that hopefully can be turned on/off on a per-device basis in the future.

Some of the graphs I would like to integrate into Cacti actually do account for graphing something special when there isn't a response.

[TheWitness]

Cacti_fod,

For right now, for a remote host, it's your only choice. In version 0.9, we have already added the feature to selectively pick the availability method to utilize at the host level including "none".

In the mean time there are a few workarounds. If you leave the availability method to SNMP for the system and clear the snmp community for the device, the device will not be checked for availability.

Also, non-host based scripts are not checked. This is in place to support legacy Cacti. Therefore, your corner case will be completely addressed in the next release and has a few workarounds in the mean time.

TheWintess