I’ve been using the latest version of Apple’s operating system for about a month now and here are my views:

It’s been said by some that Leopard is Apple’s Vista-scale disaster area and is broken beyond usefulness. Well, on that side I would disagree. There are flaws and a coupe of rather major bugs but it’s not on the scale of Vista’s retrograde step.

So, what are the problems?

(1) User interface.

The dock is awful to use in its default form. It’s hard to see which applications are running. It’s merely eye-candy. Thankfully, you can change it into a 2D form which is FAR better using a command line setting.

Similarly, whoever thought that a transparent menu bar was a good idea should be put in stocks and pelted with tomatos. It will force you to find a background which will allow you to read it. Thankfully, again, there is a command line setting to change it.

Other than that (I don’t use stacks, which is the other thing people have berated, so I can’t comment on those) it’s fine. I would have liked a different default key combination to swap between “spaces” or a mini-view of the screen layout so that I could click directly on the virtual desktop I wanted but those are relatively minor things.

(2) Outright bugs.

The problems with the default version of the X server is well documented, i.e. broken cut and paste, especially with the 3 button mouse emulation (which doesn’t work at all) and a few other things. The build in the core Xorg server tree have fixed these, we’re still awaiting the Apple official update. It’s mostly an annoyance to be honest.

But rather more importantly….

The name service subsystem is REALLY knackered!

This is important if you’re going to use Leopard on the net. With every name look-up using the built-in hostname look-up system (not DNS) the likelyhood of the look-up will fail increases until you get to the point where most name look-ups will fail or give incorrect results (which is worse).

If you don’t believe me try it out; open a web browser, either Safari or Firefox will do, load up a web site which does a lot of look-ups, e.g. Facebook with a lot of apps loaded. Refresh page, repeat. Depending upon the number of apps and adverts it will probaby take 10-15 refreshes until you’ll start to see problems. Usually you’ll start seeing missing images. The mre you reload the more the problems will appear. Some of the adverts may appear as error pages. You can try closing the application and re-starting it, or try another browser or even maybe an e-mail client. The problem’s still there. If you open an xterm or terminal application and do an nslookup the failling look-ups will work perfectly but ping/traceroute etc. will pick up the wrong entry all the same. A reboot is the only fix.

So, there you have it. Other than the major name service bug, niggles really.

OK, I don’t know if this is the 547th post but my updates are sporadic at best.

So, what have I been doing over the last couple of weeks? Not a great deal really other than visiting people the last couple of weekends.

The Saturday before last I popped down to Rachel, Graham and Chris’ place for a nice, relaxing day with tea, crumpets and flapjacks.

This weekend, after drying out from the drenching I got cycling home, I drove up to Liverpool on Friday evening and was introduced to two cats by Lindsey and Andrew. Saturday was consumed by a museum visit, Christmas tree wrangling and decoration hanging. Whilst yesterday included human kite flying (almost) and a visit to a water powered cotton mill run by the National Trust before a drive back home again.

Hmm, what else? Oh yes, I’m now a proud owner of a 2.6GHz MacBook Pro, which arrived a week earlier than predicted. i.e. marginally less than a month after I ordered it rather than slightly more than a month. It’s very shiny, in a brushed aluminium sort of way and weighs the same as the iBook G4. It’s replacing, with the help of Bootcamp & Parallels, both the iBook and the Acer laptop, which I used for mobile Linux and, when down in Cornwall, as a games machine. I “merely” have to transfer all the data and applications now.

Oh, while I’m at it.. MacOS X 10.5’s great except for a few things:

1. Transparent menu bar… bad idea. What were you thinking Apple? I now have to pick a desktop background which makes the menu bar usable.
2. The X server’s partly broken.
3. Time Machine is unusable on a G4 Mac as it will take 5/6ths of the backup run pegging the CPU at 100% with zero data transfer. On a fast Intel dual core machine it’s fine as long as you don’t unplug the external drive as it will take an hour deciding that it doesn’t need to back up anything when you next plug the drive back in.
4. The application dock may be shiny but that and the glowing blobs to tell you that an application is running make it less useful.

Oh, and if you think that I’m being partisan about knocking Apple… Microsoft Genuine Disadvantage is the spawn of several devils as well.

This morning I went into work to perform a routine installation of patches on both our mail server and file server. I expected that because there were only a few patches which required a reboot it should only take about an hour. I started the process at about 10:45am…

All went well for the first set of patches and the mail server rebooted fine. The file server initially rebooted OK but needed a couple of patches in single user mode. Again, they SEEMED to go O.K… until the final reboot. The system came back up and but couldn’t fsck the UFS filesystem on a zvol device.. so I rebooted and booted with the “-r” option to reconfigure devices…. even more errors… then lots of programs crashed on start.. so I rolled back the patches and rebooted.

The system wouldn’t boot at all, other than in failsafe mode. Rebuilding the boot_cache made no difference.. but I did notice a fleeting message saying that GRUB couldn’t mount the root partition!

Hmm… So, I boot in failsafe mode again and run fsck. OH-MY-GOD! Huge numbers of duplicate blocks, corrupted directory entries and corrupted directories. After about 30 sets of fsck runs it wasn’t getting any better so I had to cut my losses and do a full re-install.

Suffice it to say that I eventually got the system back up and running. Zpool imported the data filesystems, the user home directory UFS filesystem checked out and when I NFS exported them the clients didn’t even have problems with stale file handles.

So, a job which should have taken 1 hour actually took 11. I’m off to bed in a minute without supper ‘cos I’m so tired and, as I didn’t get to Sainsbury’s, don’t have much to eat in the house anyway.

Well, it seems that I’ve not posted anything on here for nearly a month. So, what’s been happening?

Well, my parents have visited, I’ve worked lots and last Saturday I went to a party in London, popping into the Apple Store in Regents Street on the way.

The party was for John and Stephen Hillier, their 30th birthdays. I bought them a glass tankard each from “Engraving4U” on the net and had “I’m a Thirsty Something” etched onto them. The party itself was held in the events room of a pub near Borough tube station and was a jolly nice party, though there wasn’t any dancing this time.

That’s about it really. Other than my iBook is now running Leopard (MacOS 10.5) and I’ve ordered a new MacBook Pro which may arrive in early December, if I’m lucky.

At about 9am this morning the perl script which synchronises user creation and deletion on our mail server with the rest of the network went berzerk and deleted a great many of the users’ home directories (for mail storage only).

I’ve no idea what went wrong so I’ve commented out the deletion part of the script and am now restoring all the data from the central University Tivoli backup system (horrid interface) and then from the automatic back-up of the inboxes (which runs every minute) which isn’t affected by the synchronisation script.

This script has been running without any problems for 4-5 months now. Maybe it’s a NIS glitch, as it compares a previous list of users from “ypcat passwd” with the latest version. The strange thing is that if that did happen it should have re-created the directories on the next run, all be it that they would be empty. This didn’t happen.

I’ve just turned on the telly in preparation for the new series of “Have I Got News For You?” and there’s a sitcom on, “After You’ve Gone.”

Now, I’m sure I recognise the ground floor house/living room set… it the same one as used in “Two Point Four Children” about 15 years ago. Not only that but it’s VERY similar to the one used in the series “Sykes” in the early 1970s. I wonder if it is a merely redressed set from the dawn of BBC TV comedy.

That’s what I call TV heritage. Maybe someone should get it listed. 🙂

Those of you on a long phone line will no doubt be frustrated by the speed of your ADSL Max link. With my ~50db attenuation I was getting between 2Mbps and (on a good day) 2.5Mbps with my Draytek router and it wasn’t that stable at that.

When I visited my friends Chris and Meriel Gore down in Cornwall over the summer they showed me the bargin that they’d come across. They have a much longer phone line than I have and about 70db of attenuation and are, according to BT, too far away from the exchange to get ADSL. However, somehow, they persuaded BT and Demon Internet to install ADSL on their line. Having been able to get only about 0.5Mbps on their line and that being flaky they had put a lot of research into routers with long line ability.. and they found that the 2wire BT2700HG router was the bee’s knees. Not only this, but because BT are giving them away to business customers who usually want to use their own equipment they were commonplace on eBay. Now, with tricking the router into not phoning back to BT for login details (which is very simple) they’re getting about 1.5Mbps with a rock steady connection. (See this “The Scream” forum thread for details on the firmware.)

So, a couple of weeks ago, having got fed up with the Draytek’s lack of line stability, I made the jump and bought a couple of the boxes.

The result? Well, I now have an ADSL connection which is rock steady and reliably syncs at 3.5Mbps and sometimes will sync at 4.5Mbps (though it can’t hold this over-night, though it tries hard to do so even telling me that it’s keeping the line up with a signal to noise ratio margin of -2!).

One caveat, however, if you’re using Linux and want to use WPA for your wireless encryption go for the older model with a single SSID (i.e. non-BT Fusion capable) with the SBC firmware as the Linux WPA stuff hates the newer WPA2 capable firmware.

I’ve been playing Warbook, an added application on the Facebook site and it has shown to be a good parallel with a forest ecology in some respects. Here’s why I think that:

In a mature forest the large trees take so many of the resources that at ground level hardly anything can grow. Seeds that germinate just can’t get enough energy to survive. It is only when a large tree falls that the opportunity exists for the seedlings to scramble for the new space and grow.

In Warbook the “energy” a player has is determined by their income which is generated by their free land (which they have to buy) with some addition from mines (which take free land). This land resource can be scavenged by other players who can attack other players. The player being attacked can defend using their army, which they also have to buy and upgrade. The problem for the “germinating” new players is that before you can buy and keep enough land to generate enough wealth to flourish in the game and build to the point where you can attack other players (there’s a minimum size of land holding you have to have before you can attack) you have to be able to build up a large enough army with a high enough defence strength to hold onto that land. Of course, for those who got into the game early the number of attackers was far fewer and the level of attack strength for those attackers was far lower than later in the game.

The point is that now it’s almost impossible for any new players to build up the resources needed. Larger players are “harvesting” the land resources of these new players with armies which are so powerful that there is no way the new players can defend against them and both their land and their armies are being depleated at a rate far higher than they can afford to replenish them. Just like the large trees in a forest suppressing the undergrowth. The difference with Warbook, however, is that the “big trees” aren’t restricted in their area of influence and can (and do) sap the energy of any and all opposition. It doesn’t matter, therefore, if one “big tree” dies as no space is made in the ecosystem to give  an opportunity for the saplings to develop. In other words, the game rules are fully biased in favour of early adopters and new players are essentially excluded. There are no niches in the ecosystem for such players to inhabit.

For me, the game is both interesting (in an academic sense) and highly frustrating (‘cos I can’t get to the point where I can even play effectively). Oh well. 🙂

Now, let me say this first; ZFS seems like a wonderful thing. In fact, it is wonderful except for a couple of things, which makes it totally undeployable for our new server. Actually, let’s put this another way. One thing makes it impossible because the ZFS way of doing things is mutually exclusive with the way our system (and probably a huge number of other legacy systems) works.

The main bugbear is what the ZFS development team laughably call quotas. They aren’t quotas, they are merely filesystem size restraints. To get around this the developers use the “let them eat cake” mantra, “creating filesystems is easy” so create a new filesystem for each user, with a “quota” on it. This is the ZFS way.

Unfortunately, this causes a number of problems (above the fact that there’s no soft quota). Firstly, no instead of having only a few filesystems mounted you have “system mounts + number of users” mounted filesystems, which makes df a pain to use. Secondly, there’s no way of having a shared directory structure with individual users having separate file quotas within it. But finally, and this is the critical problem, each user’s home directory is now a separate NFS share.

At first look that final point doesn’t seem to be much of a worry until you look at the implications that brings. To cope with a distributed system with a large number of users the only managable way of handling NFS mounts is via an automounter. The only alternative would be to have an fstab/vfstab file holding every filesystem any user might want. In the past this has been no problem at all, for all your user home directories on a server you could just export the parent directory holding all the user home directories and put a line “users -rw,intr myserver:/disks/users” and it would work happily.

Now, with each user having a separate filesystem this breaks. The automounter will mount the parent filesystem as before but all you will see are the stub directories ready for the ZFS daughter filesystems to mount onto and there’s no way of consolidating the ZFS filesystem tree into one NFS share or rules in automount map files to be able to do sub-directory mounting.

Of course, the ZFS developers would argue that you should change the layout of your automounted filesystems to fit with the new scheme. This would mean that users’ home directories would appear directly below /home, say.

The problem here is one of legacy code, which you’ll find throughout the academic, and probably commercial world. Basically, there’s a lot of user generated code which has hard coded paths so any new system has to replicate what has gone before. (The current system here has automount map entries which map new disks to the names of old disks on machines long gone, e.g. /home/eeyore_data/ )

The ZFS developers don’t seem to see real-world problems, or maybe they don’t WANT to see them as it would make thier lives more complicated. It’s far easier to be arrogant and use the “let them eat cake” approach rather than engineer a real solution to the problem, such as actually programming a true quota system.

As it is, it seems that for our new fileserver I’m going to have to back off from ZFS and use the old software device concatenation with UFS on top, which is a right pain and not very resilient.

This morning my plan for the day consisted of lots of house work. However, having seen the weather I thought “Blow it! Who knows when there’s going to be anouth nice, warm day this year. I’m off for a drive.”

Having wondered where to go I had the idea of driving over to Ampthill Great Park and having my lunch sat looking over the Great Ouse valley towards Bedford and then having a wander around the village I grew up in. So off I toddled.

It’s amazing to me that I left there a quarter of a century ago. A time ten years longer than the period I actually lived there, but seemingly nowhere near as long. I still remember talking to Andrew Walpole on his drive in preparation for a bike ride as if it were yesterday. The brain’s a curious thing.

Anyway, back to today…

The weather did cloud over on the way, but I didn’t let that spoil things. It was still warm enough, though it did make the light levels a pain for photography. Anyway, as I said, I had a picnic on Ampthill Great Park (so called because it is a royal park, just as Winsor Great Park, though there are no royal buildings left).

The park is the location of Ampthill Castle, one of the places Catherine of Aragon, King Henry VIII’s first wife, was placed after the divorce. In honor of this, a local member of the gentry had a memorial erected in the 18th Century.

After lunch I drove over to Houghton Conquest and parked outside the Post Office which, to my surprise, was actually open on a Sunday. That’s very much a change from the “old days.” The main changes within the village over the time since I left have mostly been the addition of new housing estates. One largish one filled in the area between the High Street and the Bedford Road, which runs perpendicularly to it. The other two estates replace the two dairy farms, which used to be owned and run by the London Brick Company. The farm down Rectory Lane had been derelict for years and the last time I remember cattle being milked there was in the early 1970s.

I took about half an hour’s wander around the village before getting back to my car and driving home. A far more interesting day than doing housework! 🙂