First Come…First Serve

30 08 2010

I don’t know if this is an improvement over the old session system or not!!!

So far…the sessions have been a bit too marketing heavy for my taste. I’m an engineer…”where’s the beef!!!” In any case…here’s my thoughts on the sessions I’ve attended thus far:

SP9649 – Imagine Virtually Anything w/ NetApp, Cisco, and VMware: This was unfortunately not much more than hot air. Nothing of real substance here…only some drivel about how much one could save by utilizing such a system. Also…it was our first glance at what’s looking to be the main theme here…”IT as a service!!!” They completely glazed over some real deployment problems and security issues inherent with deploying vFilers in a cloud service provider environment.

SS1055 – Partner Track Super Session: A little more marketing here…but there were some tidbits to take from this session. One, and I’ve felt this way for a while, it’s clear to me that VMware is hot and heavy in their move toward acquiring the mindshare, resources, and capabilities to directly sell cloud computing to end users. Even if they might say this is not the case…I can’t shake the feeling that this is the case.

ALT3003 – VMware vSphere Performance and Tuning: This is the first non-marketing session that I’ve attended this year. Notes as follows: This session is based on Lab #24 in the Self-Paced labs area. Session begins by talking about Storage I/O control. Statement made that storage is best when <30ms of latency. Unfortunately… (lack of sleep is catching up to me in this session 🙂 ) SIOC control normalized latency. In vSphere 4.1 we can also set IOPs limits. Value is per disk. 2ms kernel latency should be the max you should see… "A little bit of a glitch"…can't display real-time stats for a VM…Yeah…little glitch…too bad I see it all of the time. Turn on Mem Compression Mem.MemZipEnable (Enabled by default on vSphere) They say to be mindful of NUMA…but they don't really elaborate… Swap in rate should be watched… Should force hardware assist on AMD 32-bit guests… They talk about dropping off in increasing vCPUs in VMs…but leave off the drop from 8 vCPUs. Nothing magical in the vCPU performance management… 2000ms is a problem according to this session.




Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: