Permanent Unknown Status for SNMP Monitored Switch

[murphy_slaw]

I have an HP switch which I am trying to monitor via SNMP. I've added the device to cacti, and it appears that cacti is able to get SNMP data from it:

Code: Select all

SNMP Information
System: HP J4903A ProCurve Switch 2824, revision I.08.58, ROM I.08.05 (/sw/code/build/mako(f04))
Uptime: 1472274026
Hostname: Example ProCurve 2824
Location:
Contact: root@example.com

I've added traffic graphs for each of the ports on the switch. The switch is currently the only monitored host.

The status of the device is always "Unknown". I have verified that the poller is running properly from cron, but it appears that there were no entries added to the poller cache when the graphs were created:

Code: Select all

09/13/2006 06:20:48 PM - CMDPHP: Poller[0] ERROR: Either there are no items in the cache or polling is disabled
09/13/2006 06:20:48 PM - SYSTEM STATS: Time:1.0250 Method:cmd.php Processes:1 Threads:N/A Hosts:2 HostsPerProcess:2 DataSources:0 RRDsProcessed:0

There are no RRD files in /var/www/cacti/rra, which is not surprising since the poller doesn't seem to think the device is there.

Anyone run into something like this before? I looked in the archives and saw several similar problems, but none of the offered solutions seemed to work for me.


Versions:
CentOS 4.3
MySQL-server-standard-4.1.18-0.rhel4
cacti-0.8.6h
php-4.3.9-3.15
php-snmp-4.3.9-3.15

[murphy_slaw]

More info:

When I create new graphs for the switch, the data sources come up with "HP ProCurve 2824 - Traffic" as their name - they don't include the interface number.

They also have an empty Index Value, which prevents them from working. If I go into each datasource and manually set the Index Value to the approprate interface number, the graphs start working.

Another data point: All the graphs are created without an interace number in their name. If I rename a graph, it stops working with the following error in the debug output:

Code: Select all

RRDTool Command:

/usr/bin/rrdtool graph - \
--imgformat=PNG \
--start=-86400 \
--end=-300 \
--title="HP ProCurve 2824 - Traffic (1)" \
--rigid \
--base=1000 \
--height=120 \
--width=500 \
--alt-autoscale-max \
--lower-limit=0 \
--vertical-label="bits per second" \
CDEF:cdefa=a,8,* \
AREA:cdefa#00CF00:"Inbound"  \
GPRINT:cdefa:LAST:"Current\:%8.2lf %s"  \
GPRINT:cdefa:AVERAGE:"Average\:%8.2lf %s"  \
GPRINT:cdefa:MAX:"Maximum\:%8.2lf %s\n"  \
LINE1:cdefa#002A97:"Outbound"  \
GPRINT:cdefa:LAST:"Current\:%8.2lf %s"  \
GPRINT:cdefa:AVERAGE:"Average\:%8.2lf %s"  \
GPRINT:cdefa:MAX:"Maximum\:%8.2lf %s\n"  \
COMMENT:"\n"  \
HRULE:0#FF0000:"95th Percentile"  \
COMMENT:"(0 mbit in+out)" 

RRDTool Says:

ERROR: invalid rpn expression 'a,8,*'

Changing the name back does not fix the graph.

From what I've read on the forums, this looks like the autoincrement bug, but this is a fresh install of cacti-0.8.6h with all the official patches applied prior to creating the database.

This is driving me nuts! Any assistance would be greatly appreciated.

[fmangeant]

Hi

have you installed these patches : http://cacti.net/download_patches.php ?

Another question : how did you add graphs ? The correct way is to go to your device and click "create graphs for this host".

[murphy_slaw]

Hi

have you installed these patches : http://cacti.net/download_patches.php

Yes, all patches have been applied. They were applied before I created the database.

Another question : how did you add graphs ? The correct way is to go to your device and click "create graphs for this host".

That is how I added the graphs.

Any other information I can give you to help debug this issue?

[fmangeant]

Any other information I can give you to help debug this issue?

Maybe rebuild the poller cache, and run a mysqlcheck all the Cacti tables ?

Regards,

[murphy_slaw]

I ran mysqlcheck on all of the tables, and they were all OK.

On a hunch, I restarted mysql. When it started back up, it complained that it couldn't write to its temp directory and died. Sure enough, /var/lib/mysql/tmp was missing. I restored the directory and restarted the server, and all these issues seem to be resolved.

I'm guessing that someone was doing housecleaning on the DB and did a DROP DATABASE on tmp. Not sure why it had this affect on cacti, but it seems okay now.