mainesysadmin.com

• SSH over to the ESX server required.  You’ll need sudo or root access to complete the following.
• To view the file:  cat /etc/syslog.conf
• To modify the file:  vi /etc/syslog.conf (google for vi help if required, i = insert.  esc = exit insert mode)
• Move to the bottom of the file and add the line in the next step if you want ALL the logs to be sent over to your syslog server.
• #syslog server setup
  *.*    @syslogsrv.mydomain
• The first line in the above line is commented out by the # sign, the second line tells all the log files to be sent to “syslogsrv.mydomain”.  Use an ip address for the syslog server if you want this to work when your DNS server goes down.
• service syslog restart (restarts the syslog service)
  
• esxcfg-firewall -o 514,udp,out,syslog (opens the local firewall to allow remote syslogging)
  
• esxcfg-firewall -l (reloads the new firewall configuration)

This information was stolen from vmtn.net
http://www.vmware.com/community/message.jspa?messageID=423354

Re: SNMP Information from ESX3
Posted: Sep 29, 2006 8:48 AM in response to: cklemmer
Click to reply to this topic Reply

first edit snmpd.conf

1. vi /etc/snmp/snmpd.conf

configure it to point to the management server IP address
use a community name (here it’s public)

trapsink *.*.*.*
trapcommunity public

then start snmpd service

1. service snmpd start

configure it to autostart

1. chkconfig snmpd on

test it on local machine

1. snmpwalk -v 1 -c public localhost system

test it on another system (x.x.x.x is the esx server’s IP)

1. snmpwalk -v 1 -c public x.x.x.x system

configure your management server to receive SNMP and act upon.

This article applies to VMware ESX 3.0.1 in addition to older versions 2.5.3 – 3.0

VMware ESX server has snmpd daemon that can be used to monitor performance and send snmp traps
Virtual Center monitor many performance metrics, like CPU, Memory, network and disk I/O, but it is expensive and it is limited to those metrics.
What if you need to monitor disk space on /var/log partition? and want to receive an alert when it reaches 90%; What if you want to monitor httpd on ESX and receive a message when it fails.
If you are working in an environment with many servers and a team of systems administrators, you are most likely using a system monitoring tool or at least evaluating one.

* Edit snmpd.confWhatsup pro
# vi /etc/snmp/snmpd.conf
* Configure it to point to the management server IP address and use a community name (here it’s public)
trapsink *.*.*.* trapcommunity public
* Then start snmpd service
# service snmpd start
* Configure it to autostart
# chkconfig snmpd on
* Test it on local machine
#snmpwalk -v 1 -c public localhost system
* Test it on another system (x.x.x.x is the esx server’s IP)
#snmpwalk -v 1 -c public x.x.x.x system
* Open the firewall esxcfg-firewall -o 162,tcp,in,snmp
* Go to Whatsup machine (or any management station your have like Dell Open Manage IT Assistant or Nagios)
* Compile ESX MIBs from ESX server
/usr/lib/vmware/snmp/mibs/
VMWARE-ESX-MIB.mib
VMWARE-RESOURCES-MIB.mib
VMWARE-ROOT-MIB.mib
VMWARE-SYSTEM-MIB.mib
VMWARE-TRAPS-MIB.mib
VMWARE-VMINFO-MIB.mib
* If you haven’t added the server to be monitored yet, run discovery or add device to add the server to the database
* Create monitors for SNMP and configure credentials and communities (public, private etc…)
* Open the firewall on the management station ( here it is Whatsup pro)
* Write click the machine you want to monitor and select the monitored items (ping, http, cpu, disk space… etc.)
* Configure the actions to be triggered (email, page, event log…)