Deploying Virtual machine vs conventional install

[klee]

Is it common practice for users to just deploy vSphere OVF template for production systems?

What are the advantages and drawbacks of using this method vs. installing Nagios XI packages onto Linux OS (other than having the latest version of OS/App)?

Which is the recommended/standard method?

Thanks,

-klee

[slansing]

Sure, a lot of organizations use VM's for their production systems, the reason is that they can snapshot the entire vm to save it's current setup, and essentially everything inside it. Then, there is also the ability to effectively hot-swap hardware since it is a VM. You can shrink it, or increase it depending on what you need it to do, and how much it is using.

Of course, not everyone provisions the best hardware for their ESX/Vsphere/HyperV systems, a good portion of users also manually install XI to bare metal servers which have a wealthy amount of hardware resources available to them.

[klee]

Thanks Slansing,

Then I've got a follow up question for you:

Are you aware of a "clock drift" issue which occurs when deploying vSphere / ESX OVF Template?

We're running Nagios's latest vSphere OVF Template with VMware Hypervisor ESXi 4.1.0.

Can you please advise?

Thanks,

-klee

[slansing]

I am not, are you using the 64-bit one? Did you properly set your system and PHP time? Do you have it set up with some sort of syncing application that may be doing this?

[klee]

Unfortunately, we're using the 32 bit OVF because our test machine (HP Proliant DL380 - G4) is rather old.

I have not checked into PHP time because this is not an application issue.

Time drifting is a phenomenon which has been know to occur when running a Linux guest under VMware ESX server. It's when the kernel clock drifts from that of the harware (BIOS) clock.

I've tried a myriad of synchronization methods including NTP and "hwclock -s" without success.

Does the IPtables have to be appended in order for NTP to work?

Thanks,

-klee

[slansing]

Ah, yeah I was unsure of which you were referring to. You should just have to install ntp, make sure that it is set to start when the server does, then sync it's time with a server and start the service.

[klee]

Hi Slansing,

NTP is already installed and configured in Nagios' OVF template.
==============================================================================

I've added the NTP servers to /etc/ntp/step-tickers.
'service NTP status' returns blank
'service ntpdate restart' returns 'ntpdate: Synchronizing with time server: [FAILED]'

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

I've added NTP to IPtables to no avail. I've even modified the kernel boot options by editing /boot/grub/grub.conf and adding the line “divider=10 clocksource=acpi_pm” to prevent time drift.

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

[root@nagios-test ~]# ntpq -pn
remote refid st t when poll reach delay offset jitter

69.50.219.51 66.228.59.187 3 u 16 64 377 102.564 61657.3 5166.69
70.33.227.188 192.5.41.209 2 u 8 64 377 72.945 61659.8 4936.35
65.182.224.39 198.60.22.240 2 u 1 64 377 108.715 61666.3 3222.55
162.250.145.46 139.78.135.14 2 u 65 64 377 68.386 61620.2 6990.30

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

We're currently investigating to see if this has to do with institutional firewall. I'll keep you posted...

-klee

[abrist]

If you can update ntp from the cli, you should not need to open any ports:

Code: Select all

date
ntpdate pool.ntp.org
date

[klee]

Thanks for answering, Abrist.

Yeah, I tried that already but it didn't work.

We think it might have to do with our firewalls... I'll let you know.

I appreciate your reply.

-klee

[tmcdonald]

We'll keep this open until you return.