unix proces ntp gives wrong procescount in rz20

[Former Member]

Hi admins,

I am trying to set up "ntp" process monitoring with ccm4x

I configured a ini file in /usr/sap/tmp/procmon called "P02_ABAP_30_procmon.ini"

The content of this file is:

=============================================

$PROC

1. Check if CUPS is running

cupsd USER=root CUSTOMGROUP=LINUXOSCOL

1. Check if the Name Server Cache Deamon is running

nscd USER=root CUSTOMGROUP=LINUXOSCOL

1. Check if the Time service is running

ntp USER=ntp CUSTOMGROUP=LINUXOSCOL

$

============================================

the processes are:

cups

root 15795 15751 0 Nov11 ? 00:00:00 socket://gn016p01.gn016.kpn-post.nl:9100/ 1 a01adm 0008fj8b.A01 1 /var/spool/cups/d00001-001

nscd

root 4542 1 0 Oct29 ? 00:00:27 /usr/sbin/nscd

ntpd

ntp 3708 1 0 Oct29 ? 00:00:01 /usr/sbin/ntpd -p /var/lib/ntp/var/run/ntp/ntpd.pid -u ntp -i /var/lib/ntp

In RZ20, under

=> SAP CCMS Monitor Templates

+ cia01_A01_10

+ Monitored Processes

+ ntp

[ ] ntp: Process Count 0

+ nscd

[ ] root: Process Count 1

+ cupsd

[ ] root: Process Count 1

Why is the ntp process giving a count of 0 (zero)?

The other processes do give a count of 1 (one)

[former_member66264]

from this doc: (page 81)

[http://www.sdn.sap.com/irj/scn/index?rid=/library/uuid/d724f879-0901-0010-219b-af6665b95fee&overridelayout=true]

I learn that processes can be monitord via ntp, can you widen your check to ntp?

maybe allso remove the USER part ( for testing only)

can you also post the log from

Windows:
<host>saplocprfclog; UNIX: /usr/sap/tmp/logmon

[former_member66264]

should you not monitor ntpd ?

[Former Member]

Whether i monitor ntp or ntpd, the result is zero proces count

$PROC

1. Check if CUPS is running

cupsd USER=root CUSTOMGROUP=LINUXOSCOL

1. Check if the Name Server Cache Deamon is running

nscd USER=root CUSTOMGROUP=LINUXOSCOL

1. Check if the Time service is running

ntpd USER=ntp CUSTOMGROUP=LINUXOSCOL

$

=> SAP CCMS Monitor Templates

...+ cia01_A01_10

...+ Monitored Processes

......+ ntpd

.........[ ] ntp: Process Count 0

......+ nscd

.........[ ] root: Process Count 1

......+ cupsd

.........[ ] root: Process Count 1

[Former Member]

I changed the configfile as followed.

=================================================

1. This is a example dev_proc file

2. FOR Monitoring non SAP processes

$PROC

1. Check if CUPS is running

cupsd USER=root CUSTOMGROUP=LINUXOSCOL

1. Check if the Name Server Cache Deamon is running

nscd USER=root CUSTOMGROUP=LINUXOSCOL

1. Check if the Time service is running

ntp USER=ntp CUSTOMGROUP=LINUXOSCOL

$

1. End File

=================================================

Now, the procescount is correct.

[former_member66264]

just a small follow up:

I would stil narrow down the generality:

- see if *ntp works

- if not see if ntp* works

Most commom issue is that its started as (for example) /sbin/ntp *ntp would then work , but also /sbin/ntp

problem is that if you now start tommorow a new process containt "ntp" this will also counten.

Don't forget to do the ultimate test:

kill ntp and check if you have count 0 !