Thursday, June 22, 2006

#opensolaris

One part of the OpenSolaris first anniversary (or birthday) last week was an IRC chat party.

I'm not entirely sure I'm fluent with IRC yet. I used to chat interactively 15-20 years ago, but haven't done much since. One problem is that I get interrupted or distracted quite a lot. This isn't much of a problem for email discussions, but it's rather easy to lose the thread of a chat conversation.

First couple of times I used the builtin mozilla client. Just tried the ChatZilla extension for Firefox, which is basically the mozilla client, and it works fine.

Wednesday, June 21, 2006

Solaris Install

Mark Mayo lists his top 10 Solaris Installation Annoyances.

I'm not going to try and pretend that the Solaris installer isn't anything other than bad, but consider the following:

I've done thousands of Solaris installs over the last decade or so. I can probably count the number of times I've used the interactive installer on the fingers of one hand. No, I didn't like it much either.

The fact that most admins with a clue use jumpstart means that the interactive installer has received precious little attention. Even if the interactive installer was the best thing since sliced bread, I would never use it at work. That's not to say that the interactive installer can't be improved - it can, and must be.

I've not used a modern Linux installer (say, Ubuntu) but have used RedHat, Mandrake, Suse, Fedora. And I wouldn't necessarily say they're actually much better. Sure, they might look prettier, and feel slicker, but were often fragile (installer just bombs out) and unreliable (system all messed up afterwards). The Solaris installer might be butt-ugly and dog-slow, but produced working systems if you were prepared to suffer to the end. And if people are using interactive installers professionally, they need to get with the program and adopt some automatic provisioning scheme.

1. As for why it's slow, some work has been done on this. The biggest problem is that the distribution is compressed with bzip2. It goes way quicker if you swap out bzip2 for gzip.

2,3,4. Never happen if you use jumpstart.

5. I would argue against any attempt to change the default shell to bash. (Now, if they suggested tcsh, that might be received more favourably.)

6. Err, who on earth logs in as root?

7,8. We ought to chuck vi and vim and replace 'em with emacs...

9. Urrgh. Sendmail.....

10. Secure by default has finally made it into Nevada. But automation under jumpstart would normally knock out services you don't need.

Opera 9, Solaris x86

Not only is there a new version of Opera, but it's available for Solaris x86 as well!

Hooray!

An unfair comparison:

PID USERNAME THR PRI NICE SIZE RES STATE TIME CPU COMMAND
107 ptribble 4 49 0 1024M 971M sleep 219:15 1.75% firefox-bin
25551 ptribble 1 49 0 64M 52M sleep 0:07 0.02% opera

I still prefer firefox, but that might change. Don't know why, it just seems to work for me better than other browsers. One thing that opera does have that's useful is the ability to duplicate a tab. And I am getting rather fed up of having to restart firefox every day or two to keep memory utilization within bounds on my workstation (or cpu utilization down on the XP box at home).

Thursday, June 15, 2006

Moving On

Not as newsworthy as Scoble, but I'm changing jobs soon.

I've been in my current job for about 9 months, and it's been tough. I've never really settled properly into the organization, and there have been a lot of frustrations that I haven't been able to identify properly or address. Combine that with spending between 2 and 3 hours a day commuting, and it wasn't really a situation I wanted to continue in.

Last week got fairly frantic. I saw an advert for an attractive position with a company only a few minutes from home at the beginning of the week. Sent off my CV, got called by the agency, went for interview, went back for a second interview the next day. I liked what I saw, and they obviously liked me enough to offer me the post, so I start there in about a month's time.

I'm looking forward to it. And not just the job - 2 hours of my life back every weekday is going to make a huge difference.

Wednesday, June 14, 2006

Striding Out

A year ago today, OpenSolaris was launched and started taking its first baby steps.

Over the last year, a lot has happened. More code has been released. We've had lots of lively discussions. Many projects have started. OpenSolaris distributions have been created. There's been serious progress on building a Source Code Management system that fits the needs of Sun and the community. Members of the outside community have been filing bugs and contributing fixes.

Progress in some areas has been slow, but that's fine. We've been learning how to walk before we can run, and we love Solaris too much to risk running too fast and falling over and breaking anything. But the project is gaining strength and starting to stride out more confidently.

For my own part, I've been involved in the community since the early days of the pilot. Job issues meant I had to back off on the code contribution front for a while, but I've recently managed to get some fixes put back, with more on the way. This work seems to have been picked up by ZDnet Australia and even a Sun feature.

So we've had a good year, and I feel privileged to have been allowed to be a part of it.

, here's to many more successful years!

How fast is my SCSI?

Discovered a quick way of working out at what speed SCSI devices are connected to a Sun server at.

This only works for some SCSI cards - it seems to be a bit hit and miss. I've generally had some success on my SPARC workgroup servers, but other machines give zilch.

And the actual data transfer rate is something else again. But at least you can check that things are being set up at the rated speeds.

The command is:

prtpicl -v | grep sync-speed

and you then have to work out what the devices are.

Some examples. Here's a V240 with an attached 3310 array:

:target0-sync-speed 160000
:target1-sync-speed 160000
:target2-sync-speed 160000
:target3-sync-speed 160000
:target0-sync-speed 320000
which shows the intrenal drives at Ultra160, and the external array at Ultra320.

Here's a V440 with a pair of 3120 arrays:

:target5-sync-speed 5000
:target8-sync-speed 320000
:target9-sync-speed 320000
:targeta-sync-speed 320000
:targetb-sync-speed 320000
:target5-sync-speed 5000
:target8-sync-speed 320000
:target9-sync-speed 320000
:targeta-sync-speed 320000
:targetb-sync-speed 320000
:target0-sync-speed 320000
:target1-sync-speed 320000
:target2-sync-speed 320000
:target3-sync-speed 320000
all the disks are at Ultra320, and you see an SES target on each array that's come in at 5M.

Here's an old E450:

:target0-sync-speed 40000
:target1-sync-speed 40000
:target2-sync-speed 40000
:target3-sync-speed 40000
:target6-sync-speed 20000
:target1-sync-speed 10000
which has the disks at 40M, the DVD at 20M, and the tape at 10M.

Friday, June 09, 2006

Ultra 25

So where did the Ultra 25 come from? I don't recall any product announcement or big splash.

Still, that's not much of a surprise really, as the machine doesn't look too impressive. Only a 1.34GHz chip, which is slower than the SunBlade 1500 it presumably replaces.

I also note that the operating systems it supports include

Solaris 9 (Available September 2006)

Does this mean that there's going to be a new hardware release of Solaris 9 in September?

Wednesday, June 07, 2006

That's Reliable?

I was just looking at a Yankee Group report on server reliability, as reported by Yahoo!.

Now it's nice to hear that they think Solaris is winning on reliability. I knew that :-)

However, going beyond the headline they find:
  • 3-5 failures per server per year
  • 10-19.5 hours of downtime per server per year
Ouch!

Of course, that's server uptime, not service uptime. With a decent architecture you would have some backup so that the service would be available even if a server failed. And servers do fail, no matter how good they are, or need maintenance work.

But whatever, I don't regard 99.8% availabilty as anything like good. In fact, it's terrible.

My own experience is that Solaris is pretty damn reliable. Much better than the figures quoted, at any rate. And Windows servers themselves don't seem to be too bad (although they do seem vulnerable to major corruption events which, while rare, involve significant outage), although PC networks overall seem very fragile. Linux I've found to be less robust, with older versions simply wedging and hanging regularly (something that I believe has been dramatically improved), but I suspect a lot of Linux problems are due to people believing the myth that it's free and will run on any old piece of junk hardware, and so they use junk hadware and don't manage it properly - with predictable consequences.

The other aspect of system reliability is applications and, quite frankly, application reliability is often simply not up to scratch.