– Les Laboratoires Phoenix welcomes a new managed client, at the same time as I got my two first contractual employee (with enough job to drive them for years).
– I’ll be giving a talk at ConFoo, March 12th, called ‘Massive Scalability’. Be there, its going to be a pretty good one.
– I’ve been mandated to write another article for the European edition of Linux+DVD. Deadline is in a couples days.
– I’ve started dancing classes. (No comments please ;-))
– With the wedding happening soon, we are totally swamped with stuff to do. From food tasting to getting whatever I will wear, going through hotel reservation, decoration choices… By themself, each task is quite easy to manage, but add to that the fact we are doing most of it remotely and that Catherine schedule is just crazy.

At least, this morning, I’ve got a 20 minutes break, waiting for the bus thats going to take me to Montreal – on yet another ‘business trip’.

The next few hours will see little sleep and lots of action ; More precisely I’ll be deploying lots of hardware (2 IBM SAN, 2 core servers, 2 switchs, 2 APC, 5 branchs servers – supporting up to 20 ‘leaf’/virtual servers), and then somes (3 couples of 2 systems in high redundancy (wackamole IP ‘fencing’, shared-storage through DRBD). All that will go in ‘my’ new 48U cage @Hypertec (old nortel building) to act as a demo for some clients.

Once that’s completed, the true fun start: A very big part of this infrastructure is going to be self-healing, failure resitant and high performance. We are speaking of :

This is going to be solid, scalable, fast : the holy grail of a lot of service provider that are aiming at automatization of their ‘hosting’ business. The result of a lot of planning and testing ; the cutting-edge of cloud-computing.

Which bring me to the point of speaking about new options in the kernel. I had 2.6.24-1 (yeah, I know, old kernel) and I jumped to 2.6.30.5… and there is one nice feature that I really want to try!:
* Group CPU Scheduler (grouping tasks by user_id)

1. no more free inode on server
Database server of a client is sometimes throwing out error about inodes. This morning, I’ve decided to correct the problem once and for all.

#checking inodes
root@server:/home# df -hi | grep sda
/dev/sda1 640K 606K 35K 95% /

#checking free space
root@server:/home# df -h | grep sda
/dev/sda1 10G 5.1G 4.4G 55% /

A quick check shows 4.4G of free disk space but a little bit less than 35K inodes availables. Taking into account that each file require at least one inode… this could easily become a very dangerous situation. Resolving this issue isn’t very complicated : I’ve destroyed some of the 500k idle file on the server and move some more to another partition. Easy solution for a problem that can easily crash your server.

2. Zimbra & Sphinx ; 2 deployments
One client wanted an easy to manage, web accessible and fully featured messaging & collaboration software. Zimbra [website, download] is one of the easiest open source solution to recommend in those case. It integrate with active directory, it support full text search, tagging and got a very good web interface. Kinda of remind me of MailEnable (for Windows), but a lot less costly and available on GNU/Linux. I’ve completed quite a few single/multi-node deployment of Zimbra and I’ve always been impressed by the quality of the system. Hoo, and btw, I’m a certified Zimbra consultant.

Another wanted to add full-text search capacity to an existing MySQL database. While innodb & MyISAM built-in indexes can help speed up some queries, they are restricted by the b-tree data model. The second problems in this scenario is that the data is already partitioned in multiple database / servers. This is the perfect use case for Sphinx [website] with multi-tables/multi-databases data sources.

3. U.P.C. – BarCode reader
I’ve received a shipment of 8mm barcode reader for a client. The device is to be integrated in a point-of-sale solution my corp. (Les Laboratoires Phoenix) is developing for them. The small USB Acan device [ebay search] is easy to configure (scan specific code) and to use (act as a keyboard). True plug-and-play on Microsoft Windows XP, Vista, on GNU/Linux and Mac OSX Tiger… I’m very satisfied by the device and look forward to complete the assignment.

4. More stuff
Did way more stuff than those 3 contracts today… but hey, I can’t be writing down for all to see all my cool gigs and contracts! Some of those are actually protected by NDA or are still stealth.

Here is a quite interesting article from Intel, Improved Linux* SMP Scaling: User-directed Processor Affinity, on the subject. It shows how to distribute hardware-based interrupts to specifics CPU to lower context switching. It is way more technical that this very article, but I don’t see the need to copy/paste everything. Read it, it worth it.

So, checking out a process cpu affinity is done with following syntax:

# taskset -pc ‘pid’

The processor IDs start at 0. The CPU-ID list is available with ‘cat /proc/cpuinfo‘.

Binding a process ID (pID) to a specific processor is done with the following command. The only difference between the list/bind command is the processor id list added before the pID. It can be inserted as a single digit (1) or a list (1,2,3) or a suite (1-3).

# taskset -pc 1 ‘pid’

There you go: your now able to bind this CPU intensive oracle process on specific core on this 16-ways system.

Links : old ‘for coder’ IBM article – another one for VMWare.

W: GPG error: http://ppa.launchpad.net jaunty Release: The following signatures couldn’t be verified because the public key is not available: NO_PUBKEY 60D11217247D1CFF
W: You may want to run apt-get update to correct these problems

I always forget the GPG retrieve command (must be on one line):

#gpg –keyserver hkp://subkeys.pgp.net –recv-keys 60D11217247D1CFF

and the apt-key import command:

#gpg –export –armor 60D11217247D1CFF | sudo apt-key add -

There you go, no more problems.

On the topic of names; This is where I insert this familiar rant. I’ve never quite understood the MD as ‘multiple device’ name. Linux kernel device names, other than the ones directly linked to hardware, are normally named after their function not a ‘source’. In this instance, MD as ‘meta-device’ would make perfect sense – A device about devices. Anyway… it wasn’t named like that…

Anyway, no labs planned for any of those two techs (yet) or for the migration of md toward dm. We never know, I’ll keep you guys informed.