Home > Blogs > VMware vSphere Blog > Monthly Archives: February 2010

Monthly Archives: February 2010

SRM 4.0.1 is now available!

Hello all,

Last night we released our first patch for SRM 4.0.  It is an important one and I wanted to talk a little about it.  Incidentally, the URL for the release notes is not shown in our SRM documentation page yet but hopefully soon will be.  The readme is available at the same place where the updated bits are in the SRM downloads area – make sure to read it before applying.

Some of the important things this update covers off includes:

  • It is a complete update, meaning that it is not a collection of files that are copied over, but rather a complete install package.  This means if you need to uninstall it, you will need to also restore the database, and install the previous version to match the database.
  • Test recovery times have improved for ESX 4.0.1 hosts that use iSCSI.
  • Windows 7 and Windows 2008 R2 IP customization is now supported.
  • A problem that would cause a recovery plan to hang while powering off virtual machines at the protected site if the virtual machine's storage goes offline while the plan is running has been fixed.
  • IP customization for the /etc/hosts file will now work properly.

These are not the only fixes / improvements, but they are key ones that I know people are waiting for!

I did have a small issue when I tested the GA version (build 236215) of this patch.  On one of my lab VC's it had trouble during the update with registering the updated SRM service.  I tried the retry option, and I restarted the update several times with no improvement.  I restarted the host, and noticed that there was no SRM service and I did the update again.  The update worked fine now.  I did a test restore after the update and it worked fine.  I recommend you do a test restore before the patch and after to make sure things go smooth.  If you have any trouble with your SRM installation, generally I recommend you solve those issues before upgrading.  If the troubles you have are solved by the update than update first.  BTW, I did pass this on to engineering so if it is more than just my lab we will know about it.

If you have registered for patch notice, the SRM patch notice has not gone out yet but should real soon.  If you don't use this service you can find it here.

Let me know in the comments how the update goes and any future things you would like to learn in this blog.

Michael

F5 Accelerates Long Distance vMotion

At VMworld last year, F5 previewed and demonstrated their Long Distance vMotion solution. Since then, F5 has worked on integrating the solution into their BIG-IP product and leverage their new WAN Optimization Module (WOM).

What does this mean for VMware customers?

vMotion over distance on a traditional WAN is constrained by bandwidth and latency. Any packet loss will further throttle throughput as TCP congestion avoidance kicks in. In short distance and bandwidth constrains how far you go and how much you can pump through the link (and how many VMs you can migrate in a unit of time).

The F5 solution does a few things—it encrypts and compresses the data (i.e. less bits to send and it’s secure!); and it optimizes the transmission so it’s less susceptible to packet loss. The BIG-IP solution also automatically redirects the session to the new site.

The end result is a huge acceleration and increased success rate for Long Distance vMotions. Nojan published a table of results on the F5 DevCentral blog that showed a 3.1x to 4.7x improvement in time. e.g. vMotion time over a 1Gbps link with 20ms RTT dropped from 2:38 to 0:38. Another test over a tiny T3 link with 100ms RTT (that’s a loooonnng way) dropped from 13:43 to 3:35 with an increase in reliability of vMotion completions from <50% to 100%.

More info

F5 has produced a raft of information to help you understand and implement the solution.