Archive for cdwan
Cluster building
With all of Bioteam’s talk about cloud computing, I was a bit surprised to find myself building an honest-to-goodness non-cloud, non-virtual compute cluster a couple of weeks ago. There were wires, blinking lights, whirring fans, circuit breakers, and all sorts of messy real-world details to contend with. The servers were heavy and unwieldy, and I [...]
Wrap up on Haiti
I’ve returned to something approximating a normal routine after the trip to Haiti. I’ll admit to the usual disappointment that “everything,” didn’t change. By no means did we fix the country, or even come close (not that such was ever our intent). As on previous trips, we helped a few individuals who needed help. Most [...]
Safely back to the Dominican
We said our goodbyes this morning, gave away most of our supplies (blankets, sleeping pads, spare clothing, water bottles, etc) to the families still sleeping outdoors in the neighborhood near the clinic, and made our way to the border with the Dominican. We drove the FHM vans all the way through to the Dominican Republic, [...]
Thursday clinic
Another interesting day is wrapping up down here. We had yet another record breaking day – with 315 patients. This is an all time single day record for Family Health Ministries. Perhaps more interesting, we broke the kilo-patient barrier with our week’s total well over 1,000. We’re making do with supplies. Our group leader found [...]
State of the city
We just returned from a drive around Cite Soleil and downtown. The devastation there is much, much worse than what we’ve been seeing around the clinic. We’ve basically been in a suburb – the houses are spaced a bit further apart and two story buildings were rare. Therefore, while the devastation is very real – [...]
Tuesday, in Haiti
This will be a quick update, because I suspect that I would be asleep long before I could even summarize our entire trip. We’re in the ‘Blanchard’ or ‘Terre Noir’ neighborhood, perhaps a mile north of Cite Soleil – in Port au Prince. The clinic building where we are working survived the earthquake in remarkably [...]
Supplies
I write this from JFK airport in New York, having successfully gotten out of the house at 5am carrying four duffel bags containing just under 200lbs of (mostly) donated supplies. Our clothing and personal gear are in our carry on luggage. I remain overwhelmed by the deep vein of compassion and generosity we have tapped. [...]
Supporting Haiti
On Friday the 29th of January, I will be taking a week’s leave of absence from my day to day responsibilities at Bioteam and traveling to Haiti with a medical relief team. I’ll be going with a group called Family Health Ministries. We’ll go overland through the Dominican Republic to a clinic where my wife [...]
Cluster user training
On the same trip to India, I found myself in the position of delivering a training session to a room with incredibly varied backgrounds and skill levels. Some folks had never edited a file on the Unix command line before. Others just wanted to know where MPICH was installed. This set of slides goes from [...]
Inquiry Cluster Admin Slides
I recently took a trip to Kolkata, India to train users and administrators of an Apple / Inquiry compute cluster there. The slides I developed for that trip may be of general interest, and so I’m sharing them here. The intent of these slides is to provide an overview of the administrative tools and systems [...]
Instead of spending 20 minutes to format an ext3 volume every time you need one, you could just snapshot a ...
Let me add a few remarks about Ext3 as well.
The reason we usually recommend Ext3 instead of XFS is because ...
I'm using these slides to tune SGE 6.2 @ European Space Agency, thanks a lot! 
Great post.
Can you also try our native scale out NFS protocol? Next 3.1 release of Gluster will include it ...
So far my mental architecture is shaping up to be something like this:
Use the ephemeral disks for scratch space and ...
This shows that on-demand HPC still has a ways to go, it is very unfortunate that only ext3 is supported ...
We have had sge62u3 working fine in a 10.5 cluster using cron to startup the sge_execd process on clients. Trying ...
Instead of increasing your disk image size, you could have decreased the memory by writing a very short C program ...