Monitoring multiple Linux Processes
Hello,
I'm trying to monitor multiple processes on Linux servers. I would like to use Page/subpage like described in theHobbit-clients.cfg file, but it will only show one processes on the page for each server. Here is a what I have in the file.
PAGE=infrastructure/noc PROC batchman PROC jobman PROC netman
Infrastructure is the page Noc is the subpage.
XYMON only finds the first PROC.
Steve Newton PHH Corporation Desk 856-917-7700 Cell 856-425-1092
In <3B9C50B11919A24884D4983669EE5D7314A4D5F4 at EXMBXBG103.mtg.home.ads> "Newton, Steve (MBS)" <Steve.Newton at PHH.com> writes:
I'm trying to monitor multiple processes on Linux servers.I would like to use Page/subpage like described in the Hobbit-clients.cfg file, but it will only show one processes on the page for each server. Here is a what I have in the file.
PAGE=infrastructure/noc PROC batchman PROC jobman PROC netman
Infrastructure is the page Noc is the subpage.
XYMON only finds the first PROC.
I'm not sure what you mean by that:
Does it only check for the first of those processes, so you have only one red/green "dot" on the detailed "procs" status display? There should be one red/green dot for each of the processes listed, e.g. like that on the demo site - see http://www.xymon.com/xymon-cgi/svcstatus.sh?HOST=voodoo.hswn.dk&SERVICE=proc...
Or does it check all of the processes, but only on the first host on the page ?
Regards, Henrik
What he is asking for is if there is one proc column for each proc. There is not.
The proc column for the host includes EVERY ps. If any one ps goes down the color changes for that one column (proc). You would need to make a script to monitor a ps and create another column.
Josh Luthman Office: 937-552-2340 Direct: 937-552-2343 1100 Wayne St Suite 1337 Troy, OH 45373
On Tue, Feb 1, 2011 at 4:04 PM, Henrik Størner <henrik at hswn.dk> wrote:
In <3B9C50B11919A24884D4983669EE5D7314A4D5F4 at EXMBXBG103.mtg.home.ads> "Newton, Steve (MBS)" <Steve.Newton at PHH.com> writes:
I'm trying to monitor multiple processes on Linux servers.I would like to use Page/subpage like described in the Hobbit-clients.cfg file, but it will only show one processes on the page
for
each server. Here is a what I have in the file.
PAGE=infrastructure/noc PROC batchman PROC jobman PROC netman
Infrastructure is the page Noc is the subpage.
XYMON only finds the first PROC.
I'm not sure what you mean by that:
- Does it only check for the first of those processes, so you have only one red/green "dot" on the detailed "procs" status display? There should be one red/green dot for each of the processes listed, e.g. like that on the demo site - see
http://www.xymon.com/xymon-cgi/svcstatus.sh?HOST=voodoo.hswn.dk&SERVICE=proc...
- Or does it check all of the processes, but only on the first host on the page ?
Regards, Henrik
To unsubscribe from the xymon list, send an e-mail to xymon-unsubscribe at xymon.com
This is what I see when I list the host servers:
HOST=Hostname1,Hostname2,Hostname3,Hostname4,Hostname5,Hostname6 PROC netman PROC batchman PROC jobman
Results:
green netman (found 1, req. 1 or more) green batchman (found 1, req. 1 or more) green jobman (found 1, req. 1 or more)
This is what I see when I use PAGE/SUBPAGE
PAGE=infrastructure/noc PROC batchman PROC jobman PROC netman
Results:
green batchman (found 1, req. 1 or more)
From: Josh Luthman [mailto:josh at imaginenetworksllc.com] Sent: Tuesday, February 01, 2011 4:09 PM To: xymon at xymon.com Subject: Re: [xymon] Monitoring multiple Linux Processes
What he is asking for is if there is one proc column for each proc. There is not.
The proc column for the host includes EVERY ps. If any one ps goes down the color changes for that one column (proc). You would need to make a script to monitor a ps and create another column.
Josh Luthman Office: 937-552-2340 Direct: 937-552-2343 1100 Wayne St Suite 1337 Troy, OH 45373
On Tue, Feb 1, 2011 at 4:04 PM, Henrik Størner <henrik at hswn.dk<mailto:henrik at hswn.dk>> wrote: In <3B9C50B11919A24884D4983669EE5D7314A4D5F4 at EXMBXBG103.mtg.home.ads> "Newton, Steve (MBS)" <Steve.Newton at PHH.com> writes:
I'm trying to monitor multiple processes on Linux servers.I would like to use Page/subpage like described in the Hobbit-clients.cfg file, but it will only show one processes on the page for each server. Here is a what I have in the file.
PAGE=infrastructure/noc PROC batchman PROC jobman PROC netman
Infrastructure is the page Noc is the subpage.
XYMON only finds the first PROC.
I'm not sure what you mean by that:
Does it only check for the first of those processes, so you have only one red/green "dot" on the detailed "procs" status display? There should be one red/green dot for each of the processes listed, e.g. like that on the demo site - see http://www.xymon.com/xymon-cgi/svcstatus.sh?HOST=voodoo.hswn.dk&SERVICE=proc...
Or does it check all of the processes, but only on the first host on the page ?
Regards, Henrik
To unsubscribe from the xymon list, send an e-mail to xymon-unsubscribe at xymon.com<mailto:xymon-unsubscribe at xymon.com>
In <3B9C50B11919A24884D4983669EE5D7314A4D601 at EXMBXBG103.mtg.home.ads> "Newton, Steve (MBS)" <Steve.Newton at PHH.com> writes:
This is what I see when I list the host servers:
HOST=Hostname1,Hostname2,Hostname3,Hostname4,Hostname5,Hostname6 PROC netman PROC batchman PROC jobman
Results:
green netman (found 1, req. 1 or more) green batchman (found 1, req. 1 or more) green jobman (found 1, req. 1 or more)
OK, this is fine and what you should expect.
This is what I see when I use PAGE/SUBPAGE
PAGE=infrastructure/noc PROC batchman PROC jobman PROC netman
Results:
green batchman (found 1, req. 1 or more)
And this is obviously not correct.
I've tried reproducing your setup on my 4.3.0-RC1 installation, but I cannot make it fail this way. So I need some more information, and I'd like you to do some tests with the configuration that is NOT working (the one where you have PAGE=infrastructure/noc).
What Xymon version are you running ?
Please run this command and send me the result: xymon 127.0.0.1 "xymondboard host=Hostname1 test=info fields=hostname,BBH_PAGEPATH,BBH_ALLPAGEPATHS" (If you are running a Xymon version prior to 4.3.0-beta3, the command will be "bb" instead of "xymon", and the first word inside the double quotes must be "hobbitdboard" instead of "xymondboard").
If you look at the "info" status for one of these hosts, you'll see a "Page/Subpage" entry which is a link to a webpage. What does the URL in this link look like - specifically, does it include "infrastructure/noc" or is it something else ?
Try running the command xymoncmd xymond_client --test (that's "bbcmd hobbitd_client --test" in older versions). At the Hostname prompt, enter one of the failing hosts. Hit <enter> at the Hosttype prompt, enter "proc" as the Test, and just hit <enter> at the "ps command string" prompt. This should result in the three processes being listed as "red" - does it ?
Regards, Henrik
What Xymon version are you running ? 4.3.0.3.beta2
Please run this command and send me the result: xymon 127.0.0.1 "xymondboard host=Hostname1 test=info fields=hostname,BBH_PAGEPATH,BBH_ALLPAGEPATHS" (If you are running a Xymon version prior to 4.3.0-beta3, the command will be "bb" instead of "xymon", and the first word inside the double quotes must be "hobbitdboard" instead of "xymondboard").
[root at vmlb1mon1p bin]# ./bb 127.0.0.1 "hobbitdboard host=lb1etl1p test=info field=hostname,BBH_PAGEPATH,BBH_ALLPAGEPATHS" lb1etl1p.mortgagesvcs.com|info|green||0|0|0|0|0||0|
If you look at the "info" status for one of these hosts, you'll see a "Page/Subpage" entry which is a link to a webpage. What does the URL in this link look like - specifically, does it include "infrastructure/noc" or is it something else ?
Infrastructure Services/Enterprise Operations
Try running the command xymoncmd xymond_client --test (that's "bbcmd hobbitd_client --test" in older versions). At the Hostname prompt, enter one of the failing hosts. Hit <enter> at the Hosttype prompt, enter "proc" as the Test, and just hit <enter> at the "ps command string" prompt. This should result in the three processes being listed as "red" - does it ?
[root at vmlb1mon1p bin]# ./bbcmd hobbitd_client --test 2011-02-02 12:50:54 Using default environment file /home/xymon/server/etc/hobbitserver.cfg Hostname (.=end, ?=dump, !=reload) []: lb1etl1p.mortgagesvcs.com Hosttype []: Test (cpu, mem, disk, proc, log, port): proc To read 'ps' data from a file, enter '@FILENAME' at the prompt ps command string: Process batchman color red: Count=0, min=1, max=-1 Hostname (.=end, ?=dump, !=reload) [lb1etl1p.mortgagesvcs.com]:
Thanks
Steve Newton
-----Original Message----- From: Henrik Størner [mailto:henrik at hswn.dk] Sent: Wednesday, February 02, 2011 9:14 AM To: xymon at xymon.com Subject: Re: [xymon] Monitoring multiple Linux Processes
In <3B9C50B11919A24884D4983669EE5D7314A4D601 at EXMBXBG103.mtg.home.ads> "Newton, Steve (MBS)" <Steve.Newton at PHH.com> writes:
This is what I see when I list the host servers:
HOST=Hostname1,Hostname2,Hostname3,Hostname4,Hostname5,Hostname6 PROC netman PROC batchman PROC jobman
Results:
green netman (found 1, req. 1 or more) green batchman (found 1, req. 1 or more) green jobman (found 1, req. 1 or more)
OK, this is fine and what you should expect.
This is what I see when I use PAGE/SUBPAGE
PAGE=infrastructure/noc PROC batchman PROC jobman PROC netman
Results:
green batchman (found 1, req. 1 or more)
And this is obviously not correct.
I've tried reproducing your setup on my 4.3.0-RC1 installation, but I cannot make it fail this way. So I need some more information, and I'd like you to do some tests with the configuration that is NOT working (the one where you have PAGE=infrastructure/noc).
What Xymon version are you running ?
Please run this command and send me the result: xymon 127.0.0.1 "xymondboard host=Hostname1 test=info fields=hostname,BBH_PAGEPATH,BBH_ALLPAGEPATHS" (If you are running a Xymon version prior to 4.3.0-beta3, the command will be "bb" instead of "xymon", and the first word inside the double quotes must be "hobbitdboard" instead of "xymondboard").
If you look at the "info" status for one of these hosts, you'll see a "Page/Subpage" entry which is a link to a webpage. What does the URL in this link look like - specifically, does it include "infrastructure/noc" or is it something else ?
Try running the command xymoncmd xymond_client --test (that's "bbcmd hobbitd_client --test" in older versions). At the Hostname prompt, enter one of the failing hosts. Hit <enter> at the Hosttype prompt, enter "proc" as the Test, and just hit <enter> at the "ps command string" prompt. This should result in the three processes being listed as "red" - does it ?
Regards, Henrik
To unsubscribe from the xymon list, send an e-mail to xymon-unsubscribe at xymon.com
In <3B9C50B11919A24884D4983669EE5D7314A4D60B at EXMBXBG103.mtg.home.ads> "Newton, Steve (MBS)" <Steve.Newton at PHH.com> writes:
- What Xymon version are you running ? 4.3.0.3.beta2
It would be interesting to know if this problem also happens with 4.3.0 RC1, but I realize this upgrade might require some planning if it is a production system.
- Please run this command and send me the result: xymon 127.0.0.1 "xymondboard host=Hostname1 test=info fields=hostname,BBH_PAGEPATH,BBH_ALLPAGEPATHS"
[root at vmlb1mon1p bin]# ./bb 127.0.0.1 "hobbitdboard host=lb1etl1p test=info field=hostname,BBH_PAGEPATH,BBH_ALLPAGEPATHS" lb1etl1p.mortgagesvcs.com|info|green||0|0|0|0|0||0|
Not quite what I was looking for - you've written "field" instead of "fields" (plural) in the command. What I was after was to see if these hosts are present on more than one page in your bb-hosts file.
- Try running the command xymoncmd xymond_client --test [root at vmlb1mon1p bin]# ./bbcmd hobbitd_client --test Process batchman color red: Count=0, min=1, max=-1 Hostname (.=end,= > ?=dump, !=reload)
OK, so at least this is consistent with the "procs" status you see.
The only reason I can think of is if the hosts are present on more than one page in the bb-hosts file - then they might have a different "PAGE" value than you think. There was a change done so that PAGE will now match any of the pages a host is listed on, and if memory serves me right then this was done between beta-2 and beta-3. So that could be the problem you're having.
Regards, Henrik
On Wednesday, 2 February 2011 16:14:15 Henrik "Størner" wrote:
- Please run this command and send me the result: xymon 127.0.0.1 "xymondboard host=Hostname1 test=info fields=hostname,BBH_PAGEPATH,BBH_ALLPAGEPATHS
These fields aren't documented anywhere! A project I was working on a while back would have progressed much further if these were documented (I didn't get around to hacking source ...).
Regards, Buchan
In <201102022301.34155.bgmilne at staff.telkomsa.net> Buchan Milne <bgmilne at staff.telkomsa.net> writes:
On Wednesday, 2 February 2011 16:14:15 Henrik "St=F8rner" wrote:
- Please run this command and send me the result: xymon 127.0.0.1 "xymondboard host=3DHostname1 test=3Dinfo fields=3Dhostname,BBH_PAGEPATH,BBH_ALLPAGEPATHS
These fields aren't documented anywhere! A project I was working on a while back would have progressed much further if these were documented (I didn't get around to hacking source ...).
OK, how about this? Will add it to the 4.3.0 docs.
http://www.xymon.com/~henrik/xymon-xmh.5.html
Regards, Henrik
On Friday, 4 February 2011 00:09:17 Henrik "Størner" wrote:
In <201102022301.34155.bgmilne at staff.telkomsa.net> Buchan Milne <bgmilne at staff.telkomsa.net> writes:
On Wednesday, 2 February 2011 16:14:15 Henrik "St=F8rner" wrote:
- Please run this command and send me the result: xymon 127.0.0.1 "xymondboard host=3DHostname1 test=3Dinfo
fields=3Dhostname,BBH_PAGEPATH,BBH_ALLPAGEPATHS
These fields aren't documented anywhere! A project I was working on a while back would have progressed much further if these were documented (I didn't get around to hacking source ...).
OK, how about this? Will add it to the 4.3.0 docs.
Some very useful info there, thanks.
Regards, Buchan
participants (4)
-
bgmilne＠staff.telkomsa.net
-
henrik＠hswn.dk
-
josh＠imaginenetworksllc.com
-
Steve.Newton＠PHH.com