Tuning ESXi (6.0) logging

Rhttpproxy flood with verbose messages

Login to ESX via SSH (enable that if needed).

[root@esx1:~] cd /etc/init.d
[root@esx1:/etc/init.d] ls
DCUI                epd                 nfsgssd             sensord             usbarbitrator       vsanObserver
ESXShell            hostd               nscd                sfcbd               vmfstraced          vsandevicemonitord
SSH                 iofiltervpd         ntpd                sfcbd-watchdog      vmsyslogd           vsantraced
cdp                 lacp                osfsd               slpd                vmtoolsd            vsanvpd
clomd               lbtd                pcscd               smartd              vmware-fdm          vvold
cmmdsd              lwsmd               rabbitmqproxy       snmpd               vobd                wsman
dcbd                memscrubd           rhttpproxy          storageRM           vprobed             xorg
ddecomd             nfcd                sdrsInjector        swapobjd            vpxa

As you see, rhttpproxy has it's own initscript. Looking into it, a config file /etc/vmware/rhttpproxy/config.xml discovered. Edit it and replace verbose level to error level as shown here:

      <!-- default log level -->

Restart service by:

[root@esx1:/etc/init.d] ./rhttpproxy restart
watchdog-rhttpproxy: Terminating watchdog process with PID 34267
VMware HTTP reverse proxy stopped.
VMware HTTP reverse proxy started.

Repeat for other ESXes.

Hostd flood with warning messages

Although it is possible to fix staff in file (/etc/vmware/hostd/config.xml) exactly as we did for rhttpproxy and restart service by ./hostd restart, these parameters are avaliable via Advanced System Settings in vSphere Web Cient.

Open client, select ESX, go to Manage tab, then Settings and finally Advanced System Settings. Enter hostagent in search field, wait for results. You should select Config.HostAgent.log.level property and edit it (click on pen icon) In pop-up window switch level to error.

Repeat for other ESXes.

The following message not directly related to HostAgent, it forwards form some facility, named LikewiseGetDomainJoinInfo. Still not found the source of problem.

Hostd: [LikewiseGetDomainJoinInfo:355] QueryInformation(): ERROR_FILE_NOT_FOUND (2/0):

hostd-probe flood with verbose messages

hostd-probe woken up by crond, therefore no restart needed after fixing configuration file /etc/vmware/hostd/probe-config.xml

Replace default log level to error.

Fault Domain Manager (FDM) verbose messages

After upgrading to ESX 6.5, a new flood comes from FDM. The treatment is similar. Examine /etc/init.d/vmware-fdm to discover /etc/opt/vmware/fdm/fdm.cfg configuration file. Replace "verbose" log level to other value (I've changed to "warning") and restart service:

[root@esx1:~] /etc/init.d/vmware-fdm restart
Stopping vmware-fdm:Setting the memory limit for fdm resource pool on this host to
    0 MB

Starting vmware-fdm:Setting the memory limit for fdm resource pool on this host to
        200 MB

Updated on Wed Oct 3 13:40:09 IDT 2018 More documentations here