May 16, 2013
Caveats of Benchmarking
Benchmarking is actually something that needs to be considered very carefully and objectively. Not all benchmarks are equal. Phoronix test suite was good in the sense that you can benchmark certain workloads and I chose to focus on apache and PostgreSQL tests in the product as this closely represented the workloads I needed to improve performance on. At this point it was decided to use the PTS Desktop Live (http://www.phoronix-test-suite.com/?k=pts_desktop_live) as it was felt that this test would assure all things would be equal, regardless of platform.
Not everything was exactly equal though – none of the hardware was like for like in specifications. As I was working towards matching and improving the score of the R910 against the DL380 G5, the best I could achieve was to ensure that benchmarks were consistent. In this case the Live CD achieved this by ensuring that the same linux build and benchmarking tools were being used. I will discuss additional factors impacting the results as I continue through this series.
Read the rest of this entry »
April 19, 2013
When someone say’s their Virtual Machine is running slowly, the first thing we do is check out the performance graphs with the vSphere client. Everything looks normal, and so it is easy to dismiss it as an application issue. However, what happens when this issue is the result of a previously undetected problem that appears to have multiple components to it?
This is the problem I am currently working on now with a Dell R910 Server. The Dell R910s are quite a powerful machine designed for more intensive workloads and certainly comes as a surprise that we are seeings performance issues.
This is currently an issue being worked on, and I do not know what the outcome will be. However I will be documenting the steps being taken in order to benchmark performance and how we might improve the performance.
Read the rest of this entry »
March 13, 2013
HP released an advisory last week detailing an issue with upgrading firmware for iLO 3 interfaces of a particular firmware version to version 1.5. This appears to affects tools using the hponcfg utility, command line (CLI) interfaces and HPSUM update tools.
The good news is there is a fix for this, with HP releasing revised toolsets on the iLO 3 site, advising versions 4.1.0 of the tools or later have resolved this issue.
March 1, 2013
Where I work we have recently acquired a number of Dell R620 servers to be used in remote locations. Personally I think they are a great all round server for the light to medium workloads seen on the infrastructure we manage.
I’ve been pushing hard to use automated deployment systems that rely on the iDRAC interface, with the plan being that people take the rack and stack a previously unopened server at the remote site, use the front panel to configure iDRAC Network and then come back to work to provision the server remotely. The default username and password for the iDRAC is well known. That is until recently when we placed a server on site, confirmed we could remotely contact the iDRAC and came back, only to find the defaults not working.
It would appear that certainly the batch of servers we have, and some anecdotal reports from friends deploying Dell servers is that some server’s iDRAC interface appears to have a “faulty” default password. This was confirmed with a follow up call to Dell support. Current fix is (you guessed it) to ensure you set the password before going out on site. Whilst an inconvenience, luckily this was discovered after one server went out, and not all of them.
So just a heads for people that if you are relying on you iDRAC to remotely provision brand new boxes, you may want to set the default username/password yourself rather than making an assumption it will work.
December 10, 2012
Having recently spent a week off-site to upgrade some blade chassis and add some new hardware, one of the more pleasant aspects of the upgrades was pulling some new HP BL460 Gen8 blades out of the box and turning them on. Not so pleasant was the fact they would not get past a POST, stopping with the following enigmatic error message:
1706 – The extended BIOS Data Area in Server Memory has been Overwritten – Smart Array Interrrupt 13h BIOS Cannot conitnue – System Halted.
What was worse was that this was happening to all four blades that I had inserted. Time to call HP.
Read the rest of this entry »
December 9, 2012
Its been a busy month for me, as is always seems to be the case from about Mid-November all the way through to Christmas. My focus at the moment has been decommissioning legacy systems, of which one is particularly notable.
The last Windows Server 2003 Domain Controller in our environment gets decommissioned this week. This has been the culmination of works started in August 2009 in an effort to modernise Active Directory where I work. Upgrading AD where I have been working has not been that much of a priority, and there were both political and perception challenges to overcome as well. Some of the things that have happened as part of this work have been significant infrastructure changes including:
- Making Domain Controllers purely a single role machine.
- New DCs in sites where previously one machine acting as a DC and Fileserver
- Removal bof additional roles from domain controllers (e.g. RADIUS Auth, Certificate Authority, Scripts etc)
- New load balancing of LDAP & LDAPS connections for applications which only allowed a single authentication source
- Modification of firewall rules due to the changes in open ports in Server 2008 R2.
We got to the point of decommissioning in July 2012, but that attempt had to be aborted following some applications having unexpected issues.
But this week we are finally there. The server has been turned off for a few days to make sure there aren’t any more surprises, and this week we will shut down the server, wipe the drives and add it to the pile of decommissioned hardware.
With that done and dusted, we only need to raise the Functional level to 2008 R2 native before starting the process again with WS2012.