Xymon Mailing List Archive search

Hobbit client data parsing

9 messages in this thread

list Arnoud Post · Mon, 24 Oct 2005 10:52:29 +0200 ·
I use Hobbit client and server, but the data was not parsed propperly. This
happend 2 times now.

The disk check gives this error:
 - Filesystems NOT ok
red Expected string (Capacity and Mounted) not found in df output header

And after that the disk graphs are split up because there are 2 extra diskslices
added:
"/status"
"//dev/ad0s1b"
According to the graph the counters for these slices are "nan" instead of 0.0 or
any other number.

Does anybody have familiar problems?

Henrik, is there a smart way to make a copy of the clientdata file if there are
any alerts (red or yellow)?
list Henrik Størner · Mon, 24 Oct 2005 11:36:55 +0200 ·
quoted from Arnoud Post
On Mon, Oct 24, 2005 at 10:52:29AM +0200, Arnoud Post wrote:
I use Hobbit client and server, but the data was not parsed propperly. This
happend 2 times now.

The disk check gives this error:
 - Filesystems NOT ok
red Expected string (Capacity and Mounted) not found in df output header

And after that the disk graphs are split up because there are 2 extra diskslices
added:
"/status"
"//dev/ad0s1b"
According to the graph the counters for these slices are "nan" instead of 0.0 or
any other number.

Does anybody have familiar problems?

Henrik, is there a smart way to make a copy of the clientdata file if there are
any alerts (red or yellow)?
The entire contents of the "df" status message that Hobbit receives is
included in the "disk" status reports, after the "Expected string ...
not found" message.

I'd be interested to see what this output looks like. What kind of
systems is this ?


Regards,
Henrik
list Lars Ebeling · Mon, 24 Oct 2005 16:06:12 +0200 ·
quoted from Arnoud Post
On Mon, Oct 24, 2005 at 10:52:29AM +0200, Arnoud Post wrote:
I use Hobbit client and server, but the data was not parsed propperly. This
happend 2 times now.

The disk check gives this error:
 - Filesystems NOT ok
red Expected string (Capacity and Mounted) not found in df output header
This happens me at least once a day at the hobbitserver. I have written abou this Aug 16. I think there are timing problems when the client-script is running. This is what I got.


Tue Aug 16 13:56:37 METDST 2005 - Filesystems NOT ok
 Expected string (Capacity and Mounted) not found in df output header

System: leopg9 Tue Aug 16 13:56:37 2005
list Arnoud Post · Mon, 24 Oct 2005 23:31:01 +0200 ·
I have just made a script which makes a copy of the clientdata file when an alert is seen. Now I just have to wait and see...

#!/bin/sh
#
CP="/bin/cp"
BBHOME="/usr/local/hobbit"
DATE=`/bin/date "+%Y%M%d.%H%M%S"`
$CP $BBHOME/client/tmp/msg.txt $BBHOME/client/logs/clientdata.$DATE  
and added in hobbit-alerts.cfg:
HOST=www.foobar.com
      SCRIPT /usr/local/hobbit/server/bin/copy_clientdata.sh       
Grt,
Arnoud
quoted from Lars Ebeling


lars ebeling wrote:
On Mon, Oct 24, 2005 at 10:52:29AM +0200, Arnoud Post wrote:
I use Hobbit client and server, but the data was not parsed propperly. This
happend 2 times now.

The disk check gives this error:
 - Filesystems NOT ok
red Expected string (Capacity and Mounted) not found in df output header
This happens me at least once a day at the hobbitserver. I have written abou this Aug 16. I think there are timing problems when the client-script is running. This is what I got.


Tue Aug 16 13:56:37 METDST 2005 - Filesystems NOT ok
Expected string (Capacity and Mounted) not found in df output header

System: leopg9 Tue Aug 16 13:56:37 2005
list Arnoud Post · Tue, 25 Oct 2005 07:25:35 +0200 ·
Weird enough, the script didn't fire... but I found these errors in rrd-data.log just a few moments before I got an alert:

2005-10-25 06:13:54 RRD error updating /usr/local/hobbit/data/rrd/host.foobar.com/vmstat.rrd from 127.0.0.1: illegal attempt to update using time 1130213634 when last update time is 1130213634 (minimum one second step)
2005-10-25 06:18:56 RRD error updating /usr/local/hobbit/data/rrd/host.foobar.com/netstat.rrd from 127.0.0.1: illegal attempt to update using time 1130213936 when last update time is 1130213936 (minimum one second step)
2005-10-25 06:18:56 RRD error updating /usr/local/hobbit/data/rrd/host.foobar.com/vmstat.rrd from 127.0.0.1: illegal attempt to update using time 1130213936 when last update time is 1130213936 (minimum one second step)
2005-10-25 06:34:05 RRD error updating /usr/local/hobbit/data/rrd/host.foobar.com/netstat.rrd from 127.0.0.1: illegal attempt to update using time 1130214845 when last update time is 1130214845 (minimum one second step)
2005-10-25 06:34:05 RRD error updating /usr/local/hobbit/data/rrd/host.foobar.com/vmstat.rrd from 127.0.0.1: illegal attempt to update using time 1130214845 when last update time is 1130214845 (minimum one second step)

And some more about other graphs in the rrd-status.log file.

You just might be right Lars... seems to look like timing.
I just don't understand why my script didn't fire, this would give more certainty about the timing problem. The data should be normal.

Arnoud
quoted from Arnoud Post


Arnoud Post wrote:
I have just made a script which makes a copy of the clientdata file when an alert is seen. Now I just have to wait and see...

#!/bin/sh
#
CP="/bin/cp"
BBHOME="/usr/local/hobbit"
DATE=`/bin/date "+%Y%M%d.%H%M%S"`
$CP $BBHOME/client/tmp/msg.txt $BBHOME/client/logs/clientdata.$DATE and added in hobbit-alerts.cfg:
HOST=www.foobar.com
     SCRIPT /usr/local/hobbit/server/bin/copy_clientdata.sh      Grt,
Arnoud


lars ebeling wrote:
On Mon, Oct 24, 2005 at 10:52:29AM +0200, Arnoud Post wrote:
I use Hobbit client and server, but the data was not parsed propperly. This
happend 2 times now.

The disk check gives this error:
 - Filesystems NOT ok
red Expected string (Capacity and Mounted) not found in df output header
This happens me at least once a day at the hobbitserver. I have written abou this Aug 16. I think there are timing problems when the client-script is running. This is what I got.


Tue Aug 16 13:56:37 METDST 2005 - Filesystems NOT ok
Expected string (Capacity and Mounted) not found in df output header

System: leopg9 Tue Aug 16 13:56:37 2005
list Henrik Størner · Tue, 25 Oct 2005 07:43:24 +0200 ·
quoted from Arnoud Post
On Mon, Oct 24, 2005 at 11:31:01PM +0200, Arnoud Post wrote:
I have just made a script which makes a copy of the clientdata file when an alert is seen. Now I just have to wait and see...

#!/bin/sh
#
CP="/bin/cp"
BBHOME="/usr/local/hobbit"
DATE=`/bin/date "+%Y%M%d.%H%M%S"`
$CP $BBHOME/client/tmp/msg.txt $BBHOME/client/logs/clientdata.$DATE  
and added in hobbit-alerts.cfg:
HOST=www.foobar.com
     SCRIPT /usr/local/hobbit/server/bin/copy_clientdata.sh       
This will only work on the Hobbit server itself, since that is the
only one running the hobbitd_alert module.

And there's a recipient parameter missing from your SCRIPT command.
Should be
      SCRIPT recipientname /usr/local/hobbit/server/bin/copy_clientdata.sh       
This happens me at least once a day at the hobbitserver. I have >written abou this Aug 16. I think there are timing problems when the >client-script is running. This is what I got.
I don't see how timing can influence the client-side script. Take
a look at that script - it really is a very simple series of shell commands that redirect their output into a file. And with the
exception of the vmstat command (which puts its output into a
separate file), all of the client side commands run one at a time.

The only explanation I can find for the messed up client file you
posted back in August is if you have more than one instance of the
client running on a box - e.g. both the client that is included with
the Hobbit server installation, and the stand-alone client.


Henrik
list Henrik Størner · Tue, 25 Oct 2005 07:54:41 +0200 ·
quoted from Henrik Størner
On Tue, Oct 25, 2005 at 07:43:24AM +0200, Henrik Stoerner wrote:
This happens me at least once a day at the hobbitserver. I have 
written abou this Aug 16. I think there are timing problems when the 
client-script is running. This is what I got.
The only explanation I can find for the messed up client file you
posted back in August is if you have more than one instance of the
client running on a box - e.g. both the client that is included with
the Hobbit server installation, and the stand-alone client.
Could you try the attached version of hobbitclient.sh (goes in the
~hobbit/client/bin/ directory) instead of the 4.1.2 version ?

What this does differently is simply to build the client message in
a file named by the process-ID running the client script, so if two
clients are running simultaneously, they won't both be writing to the
same file.


Regards,
Henrik
Attachments (1)
list Lars Ebeling · Tue, 25 Oct 2005 08:06:18 +0200 ·
quoted from Henrik Størner
----- Original Message ----- 
From: "Henrik Stoerner" <user-ce4a2c883f75@xymon.invalid>
To: <user-ae9b8668bcde@xymon.invalid>
Sent: Tuesday, October 25, 2005 7:43 AM
Subject: Re: [hobbit] Hobbit client data parsing

The only explanation I can find for the messed up client file you
posted back in August is if you have more than one instance of the
client running on a box - e.g. both the client that is included with
the Hobbit server installation, and the stand-alone client.
You are right, I misunderstood the client, so I started both server and 
client.

Regards
Lars
list Larry Barber · Tue, 25 Oct 2005 09:18:07 -0400 (EDT) ·
I've noticed this happening a few times to my setup as well. It happens
when the [df] section in the clientdata is empty.

Thanks,
Larry Barber 
quoted from Arnoud Post
On Mon, 2005-10-24 at 16:31 -0500, user-714cc8ca2b55@xymon.invalid wrote:
I have just made a script which makes a copy of the clientdata file
when  an alert is seen. Now I just have to wait and see...

#!/bin/sh # CP="/bin/cp" BBHOME="/usr/local/hobbit" DATE=`/bin/date "+%Y%M%d.%H%M%S"` $CP $BBHOME/client/tmp/msg.txt $BBHOME/client/logs/clientdata.$DATE  
and added in hobbit-alerts.cfg: HOST=www.foobar.com       SCRIPT /usr/local/hobbit/server/bin/copy_clientdata.sh       
Grt, Arnoud


lars ebeling wrote:
On Mon, Oct 24, 2005 at 10:52:29AM +0200, Arnoud Post wrote: >> >>> I use Hobbit client and server, but the data was not parsed  >>> propperly. This >>> happend 2 times now. >>> >>> The disk check gives this error: >>>  - Filesystems NOT ok >>> red Expected string (Capacity and Mounted) not found in df
output  >>> header >>> > This happens me at least once a day at the hobbitserver. I have  > written abou this Aug 16. I think there are timing problems when
the  > client-script is running. This is what I got. > > > Tue Aug 16 13:56:37 METDST 2005 - Filesystems NOT ok > Expected string (Capacity and Mounted) not found in df output
header > > System: leopg9 Tue Aug 16 13:56:37 2005 >