Xymon Mailing List Archive search

New 4.3.28 build collecting no data to RRD

4 messages in this thread

list Lynn Osburn · Tue, 21 Feb 2017 17:54:49 +0000 ·
Env: HP-UX 11.31, xymon 4.3.28, gmake 4.2.1, gcc 4.2.3, rrdtool 1.6.0

New install, compiled cleanly, installed cleanly.  xymon starts ok, status page shows the xymon server as expected.

Selecting one of the attributes/tests from the server, e.g. "conn" produces the expected output, except that there is no graph. Clicking the magnifying glass icon produces the error message "Cannot access RRD directory"

/var/log/xymon  shows  alert.log, clientdata.log, history.log, hostdata.log, rrd-status.log all being appended repeatedly with message similar to
"2017-02-21 11:43:26.646604 Peer not up, flushing message queue"  (where the timestamp increases with each entry of course)

ps listing of running processes follows
=========
$ ps -ef|grep xymon
   xymon  5580  5578  0 11:44:31 ?         0:00 vmstat 300 2
   xymon  5091  5083 252 11:29:22 ?        18:51 xymond_channel --channel=status --log=/var/log/xymon/rrd-status.log xymond_rrd --rrddir=/opt/xymon/data/rrd
   xymon  5083     1  0 11:29:17 ?         0:00 /opt/xymon/server/bin/xymonlaunch --config=/opt/xymon/server/etc/tasks.cfg --env=/opt/xymon/server/etc/xymonserver.cfg --log=/v
   xymon  5090  5083 254 11:29:22 ?        18:45 xymond_channel --channel=client --log=/var/log/xymon/clientdata.log xymond_client
   xymon  5092  5083  0 11:29:22 ?         0:00 xymond_channel --channel=data --log=/var/log/xymon/rrd-data.log xymond_rrd --rrddir=/opt/xymon/data/rrd
   xymon  5089  5083 253 11:29:22 ?        16:56 xymond_channel --channel=page --log=/var/log/xymon/alert.log xymond_alert --checkpoint-file=/opt/xymon/server/tmp/alert.chk --c
   xymon  5084  5083  0 11:29:17 ?         0:00 xymond --pidfile=/var/log/xymon/xymond.pid --restart=/opt/xymon/server/tmp/xymond.chk --checkpoint-file=/opt/xymon/server/tmp/x
   xymon  5578     1  0 11:44:31 ?         0:00 sh -c vmstat 300 2 1>/opt/xymon/client/tmp/xymon_vmstat.usmkahpxbld004.5531 2>&1; mv /opt/xymon/client/tmp/xymon_vmstat.usmkahp
   xymon  5093  5083 255 11:29:22 ?        14:00 xymond_channel --channel=clichg --log=/var/log/xymon/hostdata.log xymond_hostdata
   xymon  5088  5083 254 11:29:22 ?        18:55 xymond_channel --channel=stachg --log=/var/log/xymon/history.log xymond_history --pidfile=/var/log/xymon/xymond_history.pid


listing of /opt/xymon/data directory.  Each of the listed subdirectories is empty.
=========
$ ls -l  /opt/xymon/data
drwxr-xr-x   2 xymon      xymon           96 Feb 17 11:32 acks
drwxr-xr-x   2 xymon      xymon           96 Feb 17 11:32 data
drwxr-xr-x   2 xymon      xymon           96 Feb 17 11:32 disabled
drwxr-xr-x   2 xymon      xymon           96 Feb 17 11:32 hist
drwxr-xr-x   2 xymon      xymon           96 Feb 17 11:32 histlogs
drwxr-xr-x   2 xymon      xymon           96 Feb 17 11:32 hostdata
drwxr-xr-x   2 xymon      xymon           96 Feb 17 11:32 logs
drwxr-xr-x   2 xymon      xymon           96 Feb 17 11:32 rrd
list Lynn Osburn · Tue, 21 Feb 2017 22:57:49 +0000 ·
Env: HP-UX 11.31, xymon 4.3.28, gmake 4.2.1, gcc 4.2.3, rrdtool 1.6.0

New install, compiled cleanly, installed cleanly.  xymon starts ok, status page shows the xymon server as expected.

Selecting one of the attributes/tests from the server, e.g. "conn" produces the expected output, except that there is no graph. Clicking the magnifying glass icon produces the error message "Cannot access RRD directory"

/var/log/xymon  shows  alert.log, clientdata.log, history.log, hostdata.log, rrd-status.log all being appended repeatedly with message similar to
"2017-02-21 11:43:26.646604 Peer not up, flushing message queue"  (where the timestamp increases with each entry of course)

ps listing of running processes follows
=========
$ ps -ef|grep xymon
   xymon  5580  5578  0 11:44:31 ?         0:00 vmstat 300 2
   xymon  5091  5083 252 11:29:22 ?        18:51 xymond_channel --channel=status --log=/var/log/xymon/rrd-status.log xymond_rrd --rrddir=/opt/xymon/data/rrd
   xymon  5083     1  0 11:29:17 ?         0:00 /opt/xymon/server/bin/xymonlaunch --config=/opt/xymon/server/etc/tasks.cfg --env=/opt/xymon/server/etc/xymonserver.cfg --log=/v
   xymon  5090  5083 254 11:29:22 ?        18:45 xymond_channel --channel=client --log=/var/log/xymon/clientdata.log xymond_client
   xymon  5092  5083  0 11:29:22 ?         0:00 xymond_channel --channel=data --log=/var/log/xymon/rrd-data.log xymond_rrd --rrddir=/opt/xymon/data/rrd
   xymon  5089  5083 253 11:29:22 ?        16:56 xymond_channel --channel=page --log=/var/log/xymon/alert.log xymond_alert --checkpoint-file=/opt/xymon/server/tmp/alert.chk --c
   xymon  5084  5083  0 11:29:17 ?         0:00 xymond --pidfile=/var/log/xymon/xymond.pid --restart=/opt/xymon/server/tmp/xymond.chk --checkpoint-file=/opt/xymon/server/tmp/x
   xymon  5578     1  0 11:44:31 ?         0:00 sh -c vmstat 300 2 1>/opt/xymon/client/tmp/xymon_vmstat.usmkahpxbld004.5531 2>&1; mv /opt/xymon/client/tmp/xymon_vmstat.usmkahp
   xymon  5093  5083 255 11:29:22 ?        14:00 xymond_channel --channel=clichg --log=/var/log/xymon/hostdata.log xymond_hostdata
   xymon  5088  5083 254 11:29:22 ?        18:55 xymond_channel --channel=stachg --log=/var/log/xymon/history.log xymond_history --pidfile=/var/log/xymon/xymond_history.pid


listing of /opt/xymon/data directory.  Each of the listed subdirectories is empty.
=========
$ ls -l  /opt/xymon/data
drwxr-xr-x   2 xymon      xymon           96 Feb 17 11:32 acks
drwxr-xr-x   2 xymon      xymon           96 Feb 17 11:32 data
drwxr-xr-x   2 xymon      xymon           96 Feb 17 11:32 disabled
drwxr-xr-x   2 xymon      xymon           96 Feb 17 11:32 hist
drwxr-xr-x   2 xymon      xymon           96 Feb 17 11:32 histlogs
drwxr-xr-x   2 xymon      xymon           96 Feb 17 11:32 hostdata
drwxr-xr-x   2 xymon      xymon           96 Feb 17 11:32 logs
drwxr-xr-x   2 xymon      xymon           96 Feb 17 11:32 rrd
list Erik · Wed, 22 Feb 2017 08:47:07 -0600 ·
I don't think those messages you're seeing are related to your problem.  I
have plenty of those same log entries and mine is working fine.

My first question would be permissions.  I see your listing of permissions
for the directory, but not the permissions on /opt/xymon/data, /opt/xymon,
and /opt...  what are those.  If eXecute permissions aren't set on all
parent directories, I don't think other users can read from there.

Also.. is SELinux getting in your way?  SELinux used to bite me a lot when
I first stopped disabling it and actually working with it.

Check SELinux Status:
#> getenforce

If it returns "Enforcing".. you'll have some additional work to do:

Set SELinux to "Permissive" mode.. but do NOT leave it like this:
#> setenforce 0

There are better tools for this in RHEL/Centos 7, but this is what I do to
get by in previous versions..

Make note of NUM_OF_LINES in the log file
#> LINES=$(wc -l /var/log/audit/audit.log | awk '{ print $1 }')

Next, run the Xymon GUI through it's paces..  Look at graphs,
enable/disable monitoring, acknowledge alerts, edit critical systems,
everything you can think of..

Then, generate a custom SELinux policy module using the log entries you
generated by running Xymon in permissive mode:
#> sed -n "$(($LINES + 1)),$ p" /var/log/audit/audit.log | audit2allow -M
xymon

Install the new policy module:
#> semodule -i xymon.pp

Restore SELinux to "Enforcing" mode:
#> setenforce 1
list Erik · Wed, 22 Feb 2017 16:03:36 +0000 ·
OK.  Also, httpd is what will attempt to read those files.  The user httpd
runs as (probably apache) is going to need read access to the directories
where you write your data.

On Wed, Feb 22, 2017 at 10:00 AM Lynn Osburn <user-fd907411ed25@xymon.invalid>
wrote:
Hi Erik – this is an HP-UX install, rather than any of the Linux family,
so SEL isn’t hurting me but will pursue the permissions angle to see if I
find anything useful and update the list if so.

-Lynn


*From:* Erik [mailto:user-e641b9cfddd7@xymon.invalid]
*Sent:* Wednesday, February 22, 2017 7:47 AM
*To:* xymon at xymon.com; Lynn Osburn <user-fd907411ed25@xymon.invalid>
*Subject:* Re: [Xymon] New 4.3.28 build collecting no data to RRD
quoted from Erik


I don't think those messages you're seeing are related to your problem.  I
have plenty of those same log entries and mine is working fine.


My first question would be permissions.  I see your listing of permissions
for the directory, but not the permissions on /opt/xymon/data, /opt/xymon,
and /opt...  what are those.  If eXecute permissions aren't set on all
parent directories, I don't think other users can read from there.


Also.. is SELinux getting in your way?  SELinux used to bite me a lot when
I first stopped disabling it and actually working with it.


Check SELinux Status:

#> getenforce


If it returns "Enforcing".. you'll have some additional work to do:


Set SELinux to "Permissive" mode.. but do NOT leave it like this:

#> setenforce 0


There are better tools for this in RHEL/Centos 7, but this is what I do to
get by in previous versions..


Make note of NUM_OF_LINES in the log file

#> LINES=$(wc -l /var/log/audit/audit.log | awk '{ print $1 }')


Next, run the Xymon GUI through it's paces..  Look at graphs,
enable/disable monitoring, acknowledge alerts, edit critical systems,
everything you can think of..


Then, generate a custom SELinux policy module using the log entries you
generated by running Xymon in permissive mode:

#> sed -n "$(($LINES + 1)),$ p" /var/log/audit/audit.log | audit2allow -M
xymon


Install the new policy module:

#> semodule -i xymon.pp


Restore SELinux to "Enforcing" mode:

#> setenforce 1