VIOPS: Proven Practices for Deploying and Managing VMware : Popular Threads - Logging

Community Help wanted - How to Find ESX host's CPU Usage and Memory Usage

virtualbox — Mon, 13 Apr 2009 21:00:17 GMT

Often I have been asked this question that how do you check the ESX Host's CPU and Memory Usage. We can easily check this by clicking on the host on the Virtual Center and can see the CPU Usage. Also at the same screen I can see the Memory Usage as well.

For an example I assume I am having a box of 8 CPU each 3.2 GHz. So at the Summary screen I can see that CPU Usage as : 8303 MHz.

Now for memory I can see that I have a memory of 32GB and usage is 10GB.

So the real question here is I need to know these usage (CPU and Memory) from vm-support log files. I want to replay the log file and want to know what is the usage for these two resources.

Thanks in advance.

Rg,

Prasenjit S

Community help wanted - index and analyze VI3 logs

schambers — Fri, 10 Apr 2009 23:45:00 GMT

Hello all, Texiwill, Roger Lund and Simon Shelston, and a few others, have shown great interest in helping the community get a proper grip of log management for VI3 (and of course, re-using the framework for vSphere).

Important note: anyone can use VIOPS to write their own view of how logging should be done - there is only one simple rule on VIOPS and that is all content must provide actionable guidance. Anyone can write a doc on VIOPS, there are no committees. There is a lightweight approval process that means when you write a doc on here and click "Submit for approval", then someone like me will get in touch and help you finish the doc. It's a friendly place, get stuck in!

The idea for how to document log management on VIOPS is to blend experience of our vExperts and VMware staff to produce a definitive guide to log management, as quoted on the front page of the Logging community. We want to help everyone using VI3 to be able to :

1) Understand what logs exist, where they are, their purpose - consider this a mapping exercise

2) How to manage the logs - like rotation, centralization, things like that

3) How to analyze the logs - what to watch for, what's safe to ignore, tools to use.

This will be a high quality effort if the VIOPS community pitches in... here's how you can help:

If you've ever had an issue associated with a logfile entry, and that had a resolution, then post it here

An example might be a SCSI bad code in /var/log/vmkernel - what was the code, what was the problem, how was it resolved? Is there a kb article for it?

The goal is to start at the top with the two documents below, then create a document for each log and build a red list (critical, watch for an action this!), orange list (warning, certainly capture and trend this), and black list (ignore this, developer diarrhea).

So from the doc ESX Server 3 Log Map, there are links to documents that form the red/orange/black lists for each component/log.

I'm kicking off efforts inside VMware to collate internal information, but the community input based on their experience will make this effort special and unique.

Thanks for reading, I hope you will contribute to this effort and help build the only VI3 community help index out there

Cheers - Steve

PS. Complementary to this effort is how to centralize and analyze the logs with tools like Splunk. If you have your own methods - please share. We don't need comprehensive installation details, just a high level approach for us to say "Oh THAT'S how you do it!"...

@rogerlund Talkshoe podcast on logging

schambers — Wed, 01 Apr 2009 12:20:02 GMT

http://www.talkshoe.com/talkshoe/web/talkCast.jsp?masterId=44802&cmd=tc

Check out Roger's podcast on logging, where a few folks talked about early ideas on the logging challenge and potential solutions - hat's off to Roger for starting the discussion on this big and complex topic. The challenge now, is: how does VMware, partners and vExperts simplify this topic?

To make sure we don't lose some of the fantastic discussions and high quality expertise flying around in emails, over skype, through twitter, and other channels, I have started this focused community with clear objectives of:

Say what it is that drives log management - why bother? what do you need to do for basics? who needs to see the data?
What logs are important - in VI3 there are many components, from VM down to the underlying storage array.
How to collect the logs - this might be sending remote via syslog, or log scraping to issue alerts, and what tools are available?
How to analyse the logs - who needs to see what, when and why? <-- quoting Ed Halekty here
Creating a roadmap and benchmark to improve the standards of logging out there.

Everyone has a reason for improving log management standards - admins because it just makes life easier, VMware because it enables our customers to solve more of their own challenges and reduces our costs, partners because they have another way to help their customers succeed... the list goes on.

Watch this zone for more developments - and how about getting involved? If you have a logging challenge, or you have a solution, tell us about it here by starting a discussion.

Everyone is welcome to contribute!

Steve

How to arrange log analysis lists?

schambers — Sat, 11 Apr 2009 11:35:18 GMT

Goal: to group log messages (and monitoring alerts, like SNMP traps next) into groups so you can work out:

a) CRIT What do I really have to worry about - immediate action - like a full COS root filesystem, or something that how high impact/urgency

b) WARN What messages do I need to be aware of, like a path failure or disk capacity warnings, something with medium impact/urgency

c) WATCH What messages should I track for trending - like %RDY time for a guest?

d) IGNORE What messages don't I care about?

Is there a standard already for the colour scheme for these kind of message groups? I've chosen what makes sense to me... does anyone else have input for this grouping?