Alan Hargreaves' Weblog

China Earthquake: Oh my god!

I'm supposed to be putting the finishin touches on another customer presentation this morning (in the light of one I gave yesterday). I simply had to stop doing that and get my thoughts down as I was finding it hard to focus.

Yesterday I made a comment on a colleague's blog about the earthquake, as I am also travelling in the region. I noted that I was giving a presentation to a customer at the time and actually didn't notice. I had it pointed out to me that we had had a tremor or a 'quake after I finished.

I got back to one of the offices in Beijing that afternoon and had an Australian colleague in a chat session point me at an article in an Australian newspaper about the incident mentioning a loss of life of about a hundred. This in itself was incredibly sobering, as any such loss of life is tragic.

This morning I woke up and flipped on BBC World and was utterly gobsmacked to hear 10,000 dead!

I find myself at a loss to describe my feelings. On one hand I am incredibly grateful for my own safety, but 10,000 people?

Oh my god!

This is beyond tragedy.

The loss of human life on this scale is beyond comprehension.

The China Daily lists the numbers lost in various areas. One in particluar leaps out at me. In comparison to some of the other areas the numbers are small but how can the following not tug on your heart?

Dujiangyan: Over 50 dead in a middle school. Many more are buried beneath rubble.

I almost dread going into the office today as there are certain to be people who either know that they have lost family and friends, or perhaps worse, don't know whether or not they have. My heart goes out to all of these wonderful people who have made me feel so welcome here.

I wish I knew what more to say.

Update#1

I just called my manager in Sydney to let him know that I was fine. He told me that the Australian news services are reporting on 900 kids in a collapsed school.

I am fearful that the news is only going to get worse!

Posted at 11:13AM May 13, 2008 by Alan Hargreaves in General  |  Comments[4]

Jonathan on closed MySQL extensions

I have just been reading some questions and answers with Jonathan from Tim O'Reilly. One that jumped out at me was a question he passed on from Jesse Stay. I'm going to quote both the question and answer below in full. The added emphasis is mine.

JesseStay : does he anticipate a fallout of original MySQL users or fork in the mysql code and how will they handle that if it does happen?
2008-04-25 12:26:30

JonathanSchwartz: I'm not anticipating a fork - Marten Mickos (SVP, Database Group at Sun, former CEO, MySQL) made some comments saying he was considering making available certain MySQL add-ons to MySQL Enterprise subscribers only - and as I said on stage, leaders at Sun have the autonomy to do what they think is right to maximize their business value - so long as they remember their responsibility to the corporation and all of its communities (from shareholders to developers). Not just their silo.

I think Marten got some fairly direct and immediate feedback saying the idea was a bad one - and we have no plans whatever of "hiding the ball," of keeping any technology from the community. Everything Sun delivers will be freely available, via a free and open license (either GPL, LGPL or Mozilla/CDDL), to the community.

Everything.

No exception.

I think puts things pretty much into black and white. I wonder if we will see some egg on face retractions from those who tried to pin keeping some bits proprietary on the Sun purchase, as it looks like the opposite is actually the case. That is, the Sun purchase is what is going to ensure that these extensions are open. You know, I'm not holding my breath for any "Oops I got it wrong" type comments.

Technorati Tags: MySQL, Sun

Posted at 08:34AM May 01, 2008 by Alan Hargreaves in General  |  Comments[3]

Acer wins back a customer

I guess many of you saw my rant about the poor support I was getting from Acer over getting my Ferrari 4005 fixed.

I really should have written this up earlier, as it has been resolved now for a bit over a month.

After speaking with the Escalation folks again after still having no joy, I was offered a new machine with the following specs.

• Travelmate 7720G
• 17" screen
• T5700 Core 2 duo running at 2.2GHz
• 2gb Memory
• 2x160gb disks
• webcam, more interesting looking audio, 4 usb ports, ...

A week later I had the local repairer offering me a lower spec'd machine. After I explained what I had already been offered, they agreed and we also managed to have the 3 year warranty replaced, and the internal disks were now two 250gb disks!

I've been using it for about a month now and I quite like it. Nevada simply just installed and ran. I also selected the XP option as I really don't think vista is ready yet. I've actually been using it to perform live into second life (it does have relatively nice audio and in fact it also has a line level input on the front of the machine.

While I am pleased that they replaced the machine and I am very happy with it, the fact remains that I should not have had the poor support experience in the first place.

The two things that really stand out were

1. the complete lack of correct expectation setting, especially in the light of me being obvious what my expectations were).
2. making promises that they had no intention of keeping. That is, I did not receive a single one of the promised call backs.

Folks these are Support 101 basics and really need to be fixed.

I will, however, say a big thank you to Acer for the actions that they did end up taking to address the issue.

As an aside, a little investigation of my own showed that the issue I had with the Ferrari was apparently rather common in Ferraris of that model and age, which could explain the difficulty in sourcing a motherboard. The original problem was the video adaptor dieing in such a way as to not receive a hardware notification of an event, leaving the cpu spinning on a lock, ending up getting a BSOD.

Posted at 10:31AM Feb 23, 2008 by Alan Hargreaves in General  | 

We have a fence

Well here is the result of what I've been doing this weekend. It was only possible now that our neighbor has had the retaining wall erected. There is a sense of satisfaction around completing something like this that is well outside of your experience.

The nice thing now is that we can have the dogs out the back without having to chain them up.

Posted at 04:44PM Jan 13, 2008 by Alan Hargreaves in General  | 

PSARC 2008/008 DTrace Provider for Bourne Shell

I finally got to submit the fast track for the shell provider. I've already had one comment (from Darren Reed) that I have incorporated as it made very good sense. He suggested that if we are tracking variable assignments, we should also track unset. At this point I realised that a better name for the probes would be variable-set and variable-unset. I have a working copy for SPARC with these changes now.

Below is the prefix text and the revised specification.

I am sponsoring the following fast track for myself. I am doing the
bourne shell first for two primary reasons.

1. It is the "simplest" of the shells and thus should provide the
   minimum set of probes to implement for future work in other shells,
2. Providing probes into /bin/sh gives us observability of
   approximately 60% of all of the scripts on ON.

Additionally, as it has been around for a very long time there are
quite a lot of user written scripts for it, many very badly written.

I would expect future fast tracks for other shells (eg ksh88, ksh93,
zsh, bash, ...) to reference this fast track for the minimum set of
probes.

Note the probes are currently listed as Uncommitted. As the probes
gain use I would hope to log a future fast track to increase this
stability.

A Minor release binding is initially requested. Again, once things
settle down and the interfaces stabilise it is expected that a future
case may request a patch binding.

The sh provider makes available probes that can be used to observe the
behaviour of bourne shell scripts.

Probes
------

The sh provider makes available the following probes as exports:

builtin-entry   Fires on entry to a shell builtin command.
builtin-return  Fires on return from a shell builtin command.
command-entry   Fires when the shell execs an external command.
command-return  Fires on return from an external command.
function-entry  Fires on entry into a shell function.
function-return Pires on return from a shell function.
line            Fires before commands on a particular line of code are
		executed.
subshell-entry  Fires when the shell forks a subshell.
subshell-return Fires on return from a forked subshell.
script-start    Fires before any commands in a script are executed.
script-done     Fires on script exit.
variable-set	Fires on assignment to a variable.
variable-unset	Fires when a variable is unset.

The use of non-empty module or function names in a sh* probe is
undefined at this time.

Arguments
---------

builtin-entry,
command-entry,
function-entry

	char *	args[0]	Script Name
	char *	args[1]	Builtin/Command/Function Name
	int	args[2]	Line Number
	int	args[3]	# Arguments
	char **	args[4]	Pointer to argument list

builtin-return,
command-return,
function-return

	char *	args[0]	Script Name
	char *	args[1]	Builtin/Command/Function Name
	int	args[2]	Return Value

subshell-entry

	char *	args[0]	Script Name
	pid_t	args[1]	Forked Process ID

subshell-return

	char *	args[0]	Script Name
	int	args[1]	Return Value

line

	char *	args[0]	Script Name
	int	args[1]	Line Number

script-start

	char *	args[0]	Script Name

script-done

	char *	args[0]	Script Name
	int	args[1]	Exit Value

variable-set

	char *	args[0]	Script Name
	char *	args[1]	Variable Name
	char *	args[2]	Value

variable-unset

	char *	args[0] Script Name
	char *	args[1]	Variable Name

Examples
--------

1. Catching a variable assignment

	Say we want to determine which line in the following script has
	an assignment to WatchedVar:

	#!/bin/sh

	# starting script
	WatchedVar=Value
	unset WatchedVar
	# ending script

	We could use the following script

	#!/usr/sbin/dtrace -s

	#pragma D option quiet

	sh$target:::line { self->line = arg1; }
	sh$target:::variable-set /copyinstr(arg1) == "WatchedVar"/ {
	        printf("%d: %s=%s\n", self->line, copyinstr(arg1),
		    copyinstr(arg2))
	}
	sh$target:::variable-unset /copyinstr(arg1) == "WatchedVar"/ {
	        printf("%d: unset %s\n", self->line, copyinstr(arg1)); }


	$ ./watch.d -c ./var.sh
	4: WatchedVar=Value
	5: unset WatchedVar

2. Watching the time spent in functions

	#!/usr/sbin/dtrace -s

	#pragma D option quiet

	sh$target:::function-entry { self->start = vtimestamp }
	sh$target:::function-return {
		@[copyinstr(arg1)] = quantize(vtimestamp - self->start)
	}

	Similar for the other entry/return probes, with the exception
	of subshell as the probe name is unavailable.

3. Wasted time using external functions instead of builtins

	This script is copied from the DTrace toolkit. It's function
	and how it works should be relatively self explanatory.

#!/usr/sbin/dtrace -Zs
/*
 * sh_wasted.d - measure Bourne shell elapsed times for "wasted" commands.
 *               Written for the sh DTrace provider.
 *
 * $Id: sh_wasted.d 25 2007-09-12 09:51:58Z brendan $
 *
 * USAGE: sh_wasted.d { -p PID | -c cmd }       # hit Ctrl-C to end
 *
 * This script measures "wasted" commands - those which are called externally
 * but are in fact builtins to the shell. Ever seen a script which calls
 * /usr/bin/echo needlessly? This script measures that cost.
 *
 * FIELDS:
 *              FILE            Filename of the shell or shellscript
 *              NAME            Name of call
 *              TIME            Total elapsed time for calls (us)
 *
 * IDEA: Mike Shapiro
 *
 * Filename and call names are printed if available.
 *
 * COPYRIGHT: Copyright (c) 2007 Brendan Gregg.
 *
 * CDDL HEADER START
 *
 *  The contents of this file are subject to the terms of the
 *  Common Development and Distribution License, Version 1.0 only
 *  (the "License").  You may not use this file except in compliance
 *  with the License.
 *
 *  You can obtain a copy of the license at Docs/cddl1.txt
 *  or http://www.opensolaris.org/os/licensing.
 *  See the License for the specific language governing permissions
 *  and limitations under the License.
 *
 * CDDL HEADER END
 *
 * 09-Sep-2007  Brendan Gregg   Created this.
 */

#pragma D option quiet

dtrace:::BEGIN
{
        isbuiltin["echo"] = 1;
        isbuiltin["test"] = 1;
        /* add builtins here */

        printf("Tracing... Hit Ctrl-C to end.\n");
        self->start = timestamp;
}

sh$target:::command-entry
{
        self->command = timestamp;
}

sh$target:::command-return
{
        this->elapsed = timestamp - self->command;
        this->path = copyinstr(arg1);
        this->cmd = basename(this->path);
}

sh$target:::command-return
/self->command && !isbuiltin[this->cmd]/
{
        @types_cmd[basename(copyinstr(arg0)), this->path] = sum(this->elapsed);
        self->command = 0;
}

sh$target:::command-return
/self->command/
{
        @types_wasted[basename(copyinstr(arg0)), this->path] =
            sum(this->elapsed);
        self->command = 0;
}

proc:::exit
/pid == $target/
{
        exit(0);
}

dtrace:::END
{
        this->elapsed = (timestamp - self->start) / 1000;
        printf("Script duration: %d us\n", this->elapsed);

        normalize(@types_cmd, 1000);
        printf("\nExternal command elapsed times,\n");
        printf("   %-30s %-22s %8s\n", "FILE", "NAME", "TIME(us)");
        printa("   %-30s %-22s %@8d\n", @types_cmd);

        normalize(@types_wasted, 1000);
        printf("\nWasted command elapsed times,\n");
        printf("   %-30s %-22s %8s\n", "FILE", "NAME", "TIME(us)");
        printa("   %-30s %-22s %@8d\n", @types_wasted);
}


Stability
---------

Element Name    Class           Data Class
------------------------------------------
Provider        Uncommited      Uncommited
Module          Private         Private
Function        Private         Private
Name            Uncommited      Uncommited
Arguments       Uncommited      Uncommited
------------------------------------------

Technorati Tags: Solaris, OpenSolaris, DTrace

Posted at 09:27AM Jan 09, 2008 by Alan Hargreaves in OpenSolaris  | 

Happy Birthday to me

Happy Birthday to me,
I'm now 43.

Kind of sad that that's the best rhyming couplet I can come up with.

Posted at 09:02AM Jan 09, 2008 by Alan Hargreaves in General  | 

Customer Service (?) from Acer Support

Updated twice (see end of entry)

After having my Dell notebook die outside of warranty two years ago, I made sure to buy the extended warranty when I replaced it with an Acer Ferrari 4005 in November/December 2005.

About a month ago, (around 2 years after purchase interestingly, just like the Dell), I started having problems with it. I noticed a couple of cracks on the top of the screen and that I had two keys starting to be questionable (they work about 50% of the time). Shortly after that I started seeing windows regularly crash and Solaris hang. A little investigation showed that the video card (on the main board) was starting to play up.

So, obviously I arranged to have it sent back under warranty.

This is where the tragedy of errors begins.

I got the tracking number when DHL picked it up from Gordon at about 9:30.

It looks like the driver drove all day before dropping it into his depot at 8:00pm that evening (end of shift?). It then looks like it went out of the depot for an hour at about 3am before being returned at 4am and was finally delivered to the repair center in Flemington at about 9:15am the next morning. Note the distance between the repair center is about an hours drive, if that.

OK I was called a few days later to be told that the screen would not be replaced under warranty, but if I was willing to pay for it they'd replace it. I declined. My belief at this point was that the main board had been done and it would be shipped back to me shortly, and everything I said communicated this expectation. Nothing was done to correct it.

The following week I had a call stating that they wouldn't be replacing the keyboard as there had been a "liquid spill" on both the keyboard and the disk. I reached over my desk and picked up the disk (which I kept as I did not wish it reformatted) and said "hmmm interesting, I have the disk in my hand and see no such evidence. If there is evidence of a liquid spill on the disk currently installed, it must have happened there", which of course prompted denials. The upshot was that they also were not going to replace the keyboard unless I paid for it. Given I can clean a keyboard myself I declined. Again the expectation thing, in fact I also changed the address for them to return it as I would be working from home the next few days. No correcting of that expectation.

At this point they had had it for a week.

Come Friday, I was concerned that I still hadn't got it. So, yet another call to the support line (and the interminable wait being constantly informed of my position in the queue). At this point I discover (for the first time) that the main board is actually on back-order, and everyone that I had previously spoken to had been aware of this and not passed it on. They would not tell me the expected date of delivery.

Monday I tried again after hours and got probably my first good experience with Acer Customer Service. The guy was very helpful and understanding and did actually tell me that my part was expected on November 29.

I called during the day the next day to speak with some in-hours person about my disappointment in the way that I had been mislead and got the usual platitudes.

OK, come Thursday (Nov 29), I called a bit later in the day to verify that work had either been complete or had at least started. The back-order had slipped to the next day.

Let's try again, Monday December 3. It had slipped to December 5 (the following Wednesday).

At this point I demanded that if it slipped again that I would be called immediately that information was known and received that commitment.

I was on training Tuesday & Wednesday so didn't get to call them until after hours on Wednesday. Again, I got another person who really tried their best (hmm why do I have better experiences with their after hours staff?). Unfortunately the case ticket had not been updated, but they offered to email the technician who was doing the repair so that they would call me first thing this morning.

You guessed it, it didn't happen.

I called up a few minutes ago (and waited on a queue that started at length 20). Gave my case number and Identified myself. I was then told that the part had not arrived and that there was an outstanding query to their supplier about when it would be delivered.

I was livid. Especially at the "I understand why you are upset" platitudes.

Apart from not being called this morning as per commitment, from last night, the delivery had slipped again and no-one thought to get in touch with me.

Acer Support is not a small company, but this kind of behavior makes them look decidedly mickey mouse.

I have demanded to be told the instant that they know the new delivery date, and received yet another commitment to be called this afternoon with the information.

It goes to their management if they drop the ball again, I've had it. I've been without my notebook now for more than three weeks. Maybe I should send them a bill for the time I've wasted on the phone trying to sort this out. That should come close to replacing the screen!

One thing is for sure, much as I liked the machine and the good performance it has given me, I will not be replacing this machine with another from Acer when the time comes.

A suggestion to the folks on the phones at Acer support. Your customers are your reason for existence. Without them, you would not have a job. When you make a commitment, you honor it. A Suppliers support division is one of the main reasons that that they get return custom. Here in the Sun Support office in Gordon, we used to have a poster up that simply said "It can take years to win a customer, and seconds to lose them". Truer words were never uttered. If the first people that I had spoken to had set my expectations correctly from the outset, I would not be as angry as I currently am. A simple "To fix your video card, is a main board replacement. We've had to order this in and they normally take 3-4 weeks" would have done this.

If I had treated a customer with the obvious contempt with which I feel I have been treated by Acer Support, I would expect and deserve a serious ass-kicking from my management.

Update #1

Well, it's just gone 5pm here and I have not received the promised call from Acer. Tomorrow we start talking to call center management. Sigh, I wish it hadn't come to this. Folks, you don't promise a customer something only to get them off the phone so you can forget about it.

Update #2

Just got off a phone call with their escalation department speaking with someone called Frank. Unfortunately the phone system somewhere between us was playing up. I noticed while in the queue that I had extended periods of silence and it looks like one of those occurred while I was speaking with this person.

He agrees that this has taken a long time and told me that there is still no ship date on the back-ordered main board. While I am skeptical about this, he has committed to having the repair folks actually see if the board can be repaired, as apparantly this is one of the things that the Highpoint folks in Flemington do. I am to expect to be called by them early next week.

We shall see.

Just before we got cut off, I was pointing out that there appear to me to be call-centre folks in the support centre who will commit to almost anything to get an upset customer off the phone and then ignore it. Unfortunately I did not get to hear his response to this as that's when the line went silent again and then was disconnected a minute later.

Posted at 02:08PM Dec 06, 2007 by Alan Hargreaves in General  | 

Listening to Ian Moss

I got to hear some tracks from Ian Moss' 2005 Album "Six Strings" recently.

WOW!

I was already aware of Ian and a lot of his work, but this album of him solo, unplugged is nothing short of awesome.From classics such as Tucker's Daughter through to new stuff like Message from Baghdad (featured on his myspace), Two seconds too long and All alone on a rock.

I have a new favorite album, well I will when it arrives!

The album was recorded live at a performance, I'd love to find out there is a video to this!

Posted at 09:17AM Dec 01, 2007 by Alan Hargreaves in Music  | 

Mutex Contention vs number of cpus

I've had a few cases recently that have brought this issue to the fore.

It's amazing how many people think that the answer to all performance issues is to simply throw more cpu at the problem.

Let's work through this thought (and this holds true for other queuing type locks too).

1. On Solaris, if the mutex holder is on proc, the waiter spins instead of blocking and sleeping
2. If we have sufficient threads wanting the same mutex, we can quickly utilize all cpus in a box
3. The more threads we have in the queue for a mutex, logically the longer any given thread will take to progress through this queue to obtain it.

What does this tell you about what is likely to happen if you add more cpus into the mix?

It's relatively obvious now, isn't it.

1. More cpus stuck spinning in kernel space
2. Longer mutex queues
3. Longer average time to obtain mutex for each thread that wants it.

The obvious consequence being that adding more cpus can actually have the effect of making the problem worse.

Sigh.

Posted at 11:23AM Nov 20, 2007 by Alan Hargreaves in OpenSolaris  |  Comments[2]

Making a Stomp Box

I posted this over on my myspace blog, but I'm sufficiently please with myself that I thought I'd post it here too.

Well I've just finished the main work in making a new stomp box.

I did a google search and came up with the following instructions:

2 x 350mm x 70mm (x20mm) hardwood
2 x 310mm x 70mm (x20mm) hardwood } Gives you a total of 350mm x 3500mm square, 70mm high.

For the top: 350mm x 350mm 3ply (make sure all ply sheets are same thickness, first one they gave me was actually thinner because the top and bottom sheets of ply where as thin as a card).

Then i just used a round file to make a small circular groove for the mic lead to sit because if the stompbox isnt flush with the ground all the way around...ahh... block your ears! FEEDBACK.

Well the best I could do for the sides was 66mmx19mm marante. I had Bunnings cut up the pieces to the above dimensions for me, picked up some nails and glue. Unfortunately with Bunnings, you pay for the whole bit of wood they cut up. The marante was 1.8m long (so not really a lot left over there), but the 3-ply was actually a 450x900mm cut. All up, cost me about AUS$20.

I only screwed up on one thing with it, I put the 350mmx350mm piece on upside down, so now I have a nice hole to fill before I stain and varnish it. I also still have to make the cable hole for the mike, but I'm pretty happy with it.

Posted at 08:40AM Nov 20, 2007 by Alan Hargreaves in Music  | 

Toy Recall - HOLY CRAP

I just had my dad call me to tell me about this toy recall.

The toy (Bindeez Craft Beads) won the Australian Toy of the year award! Now they find that

"The colourful beads contain the same chemical as the illicit drug GHB, otherwise known as 'fantasy' [...] the product actually contains 1,4-butanediol [...] When metabolised, 1,4-butanediol converts into GHB."

My (almost) 9 year old daughter loves playing with these things!!!

Sometimes as a parent I have the absolute sh*t scared out of me.

I will be packing these up to be returned as soon as I get home!

Posted at 08:07PM Nov 06, 2007 by Alan Hargreaves in Kids  |  Comments[2]

The official melbourne cup site can't cope?

I simply have to tongue in cheek blog this. For those that don't know, Australia stops on the first Tuesday in November for the largest horse race in the country. The Melbourne Cup. Shortly after the race finished a colleague gathered the below screen shot from the official web site.

Do you think they might need to talk to Sun Australia about the new T2 boxes ( e.g. T5220 & T5120), Solaris 10 and Our Application Server ? :-)

Posted at 04:10PM Nov 06, 2007 by Alan Hargreaves in Solaris  |  Comments[1]

Sun Developer Days

OK, I got back from CEC on Saturday a week back and walked into the house at about 9:30 absolutely knackered. About 2pm my pager went off and I discovered that I was on VOSJEC duty that weekend and ended up with a righht horror of a call that lasted the rest of the weekend (that I won't go into detail here with, save to say that I got an action plan out to these ghuys at about 00:30 on Monday morning.

Early Monday morning (ok I did get some sleep, this is real morning about 10-11), I got a call from Laurie Wong. Apparantly the DTrace speaker they had organised for the Developer Days couldn't make it and they really couldn't find anyone else. After some discussion with my boss, we agreed that I would go fly to Melbourne the next day to cover this and also cover Sydney on Wednesday.

Had an awful time actually using the system that we are supposed to use to book the flight, ended up taking me a bit over an hour and by that time the fare had risen 50% !!! Anyway got that all sorted and boy am I glad that I booked to get my self well ahead of when I spoke.

First off, I was using someone else's slides, so of course I had to work out what I was going to say to each one (I use flash cards to remind me of what I want to talk about so I'm not just reading the slides). Going through the slides I noticed that the information on the javascript provider was actually out of date. Indeed, you can actually download a firefox 3.0 alpha that has the new provider in there and looks pretty damned spiffy. So, I updated that stuff, then I discovered that there were actually two sections of the talk not present in the slides. This was the "tie it all together" bit and the summary. Well I didn't have the time to write a "tie it all together bit", so I removed that from the agenda slide and did up an "in conclusion" slide.

The other part of being glad I booked an earlier flight is that even though we boarded close to time, we were about an hour late getting off the runway! I got in to Melbourne at about midday. Fortunately we were able to put another speaker in front of me so I could finish writing the talk which I ended up giving at 4pm.

Anyway, the talk covered some background on DTrace (and the slide author provided some really nice graphics and animations), and discussion of various providers. In particular I talked about PHP, javascript, and postgresQL. I did demos for some of the basic DTrace, javascript and postgreSQL.

I Also touched on the shell provider I'm working on and encouraged folks to get involved with working on and testing new providers.

Amazingly, without having timed this or even thought about the length, I managed to finish exactly on time.

Laurie took me into the QANTAS lounge where we were able to relax a little before the flight home. With the flight and the train trip I got home about midnight.

The next day was in Sydney, so I only needed to take the train into the city.

After finding the venue (google maps on a treo 750 is really useful!), I sat in on a couple of the other talks and quite enjoyed those. In Sydney my talk was at 3:15 and again went pretty well.

Headed home after being treated to a really nice dinner at Doyle's on the Harbour.

Unfortunately I had a prior commitment on Friday so I couldn't give the talk in Canberra.

These were probably the largest audiences that I have ever presented to (combining both talks I spoke to about 580 people). I actually enjoyed it and I think my audience had a bit of fun as well. It's nice to do this kind of thing every so often.

Technorati Tags: Solaris, OpenSolaris, DTrace

Posted at 10:13AM Oct 23, 2007 by Alan Hargreaves in OpenSolaris  | 

Catching up

OK I know I've still got some catching up to do, like I haven't done the last day of CEC. However there will be a couple more blogs before I get there about interesting stuff that happened after I got back. Including the developer days I spoke at in Melbourne and Sydney last week and some stuff I learned while giving those talks.

Posted at 09:44AM Oct 23, 2007 by Alan Hargreaves in General  | 

Tuesday at CEC

I had planned for a 5:30am start to get some online stuff out of the way before breakfast. My alarm had other ideas, not waking me up until 6:30am. I lay in bed for a bit thinking my room mate was showering etc, only realizing about 20 minutes later that it wasn't him and he had wandered off somewhere. Got down to breakfast about 7:15 and came across someone else I'd hoped to meet in Martin Canoy (who helps manage the Performance V-Team). Interesting eating breakfast in a room that had a semi-trailer parked up the front with a black-box on it.

The general session started at 8am with Introductions by Dan Berg and the a talk from our VP for Eco Responsibility David Douglas. One of the standouts in this for me was the work done on the Santa-Clara Data-Center. This site is now a showcase for a green data-centre and tours are conducted through it for companies interested in just what we did. We were encouraged to get customers to tour it, but I would love to see a few you-tube clips put up about it it and perhaps a 30 minute documentary that we could give out on a DVD (are you listening guys), as the world is a lot bigger than the Bay area.

Next up we again we had Jonathan spend some time talking to us and answering some surprisingly targeted questions, all handled very well.

After a short coffee break, we came back for the release of the new T2 boxes. I have to say that these boxes look awesome and are going to kick some major ass in the marketplace. Given the launch was done in front of 3500-4000 technical people, it was observed during the question tat te type of questions that the panel were receiving were a lot more technical than they would normally get during a product launch. Many of the questions focused on the desire for a T2 based laptop or workstation, to whic Andy joking replied along the lines of "I can't comment on future products". When the workstation question came up he pointed back to "i thought we were looking at a laptop" or something like that :) One of the questions asked about a dual socket T2 based machine, and it was confirmed that this is currently being worked on. Now that will be a box to contend with!

During the break between 11 & 12, I spent some time at the Second Life booth in the pavilion, giving a second avatar in the same general area as the one they were using to show things off.

I had a whole lot of breakouts I wanted to go to today, but I also wanted to get some Solaris Certifications done. Unfortunately, the room only seated 30 people and when it opened at 12:15pm, the queue already had 45 in it. The took the sensible move of taking folks names and giving them a rough estimate of when they should come back. After getting my name on the list I wandered over to the installfest and started a live-upgrade of this notebook to nevada build 74 (as I was having some punchin problems with build 73). After I got the initial copy done and the upgrade kicked off, I went back to the certifications to find that there were only about half a dozen people in front of me on the list. I only had to wait 5-10 minutes to get in.

I was a little concerned that as I was doing these examinations cold, that I might not get through. I should not have worried. My 16 years of SA, as well as maintaining my own machines while at Sun, and doing kernel work and Open Solaris Advocacy stood me in good stead.

The Solaris 10 Admin (part one) allowed 90 minutes to complete it and had a 61% pass mark. I finished it in 30 minutes and scored 71.2%. Woo hoo. Heartened by this I spoke to a proctor and asked if I could do part 2 as I'd only used 30 minutes of the potential 90. He agreed and I started part 2. This was a little harder, but I knocked it over in 45 minutes and scored 70.5. So now I'm certified (well you knew I was certifiable, but that's another story).

Wandered back to the installfest to pick up my notebook. Ran up the new build and was pleasantly surprised to see punchin working correctly.

By the time all this had finished, the final breakout session was about to start, but there really wasn't anything I wanted to see in there. As it was now 6pm, headed back to my room to get change for the party at the Palms, as the buses were leaving at 7pm. I'm glad I did, as when I got back down to the lobby at about 6:40, there was already a very long line for the buses.

The party was a blast. The live band was awesome. There was also a number of games put on for us, like air hockey, video mountain biking, surfing, gun fighting, etc. The mountain biking game was murderous and really wore you out in a hurry, of course not being able to adjust the seat to a proper height made pedaling difficult. The game I fell in love with, though, was a water skiing one. Anyone who played it understood just why you would get to the whooping and hollering while you were doing it. It was incredible fun I lost count of the number of times I rode it!

I spent a lot of tonight looking for a couple of friends from the US that I wanted to spend some time with, but didn't manage to find them. Maybe they didn't go, maybe there was just too many people (note to self, get phone numbers next time). I did find a few Australian colleagues whom I spent time with, and Bob Sneed introduced me to some really nice friends of his whom I also spent some quality time with. Oh I also met Bela Amade from the EMEA cluster group who I have also done a lot of work with, as well as a number of other folks that I had backlined escalations with, whose names are too many to recall this late at night (sorry folks, I did enjoy meeting you, it's always good to put a face to a name).

About 10:45, they started herding us to the buses (as the last bus back would be at 11). I continues chatting with one of the folks Bob introduced me too while traveling back on the bus.

All in all a wonderful day. We've got the group specific meetings tomorrow and after that I am doing a podcast with Don Grantham, which should be fun.

I forgot to mention that on Monday at lunch I also met someone else I had ben hoping to meet. Dimitri DeWild. Dimitri works in a similar group to me in EMEA, and we have long communicated with each other over email and IM.

Technorati Tags: cec2007, suncec2007

Posted at 05:46PM Oct 10, 2007 by Alan Hargreaves in General  |