OK,

I was to give up on my original Core Duo MacBook Pro, simply due to the fact that it routinely ran out of RAM. It was maxed out to 2gb of RAM within days of purchasing it, and served its purpose well, until that purpose was extended to include running a VMware Fusion instance on a regular basis. We use several Windows development tools to create and maitain our Win HelpMate Software, and each time I build a new version I have to use VMware. Unfortunately, when I do, it's really tough to keep working in Mac OS X as well due to the fact that the OS starts swapping like crazy.

So, figuring Apple would release a new Pro notebook, I set aside some money and hoped for the best at MacWorld last week. Well, the MacBook Air is a stunning engineering acheivement. On first hearing the specs, I was immediately disappointed. No Ethernet, no optical drive, a single USB port, no Firewire (and therefore no target disk mode for troubleshooting). In order to use it, I'd have to carry a sack of dongles with me, hardly what any sysadmin/trainer/developer wants. I needed at least 4gb of RAM, 160gb of storage, gigabit ethernet, and possibly an expansion slot, not to mention at least two USB ports.

There was simply no way that this "Airbook" was going to replace my do-it-all MacBook Pro, which had replaced my G5 Desktop at my primary computer a couple of years ago.

Then I went to the Apple Booth at MacWorld, stood in line, then put my hands on a MacBook Air. That was indeed a mistake.

Because now I really want one. I know I shouldn't have one but I really can't help myself.

So, being a problem solver at heart, with money to spend, and a need, I started to re-examine the Macs in my life and how the MacBook Air might fit in.

1) The MacBook Pro would have to go. It's optical drive was long dead anyway, so it wouldn't make a good primary to the MacBook Air as secondary, even though I'd upgraded it to a 250 gig internal hard drive. It would go to eBay, a friend or family member...when "docked" at my desk at home, it's hooked to a 20 inch display that the MacBook Air could easily handle...

NEXT! (whoops, bad pun there)

2) I like the idea of the Time Capsule as a backup device for home use. But I already have a Mac Mini I use as a media center that I backup my MacBook Pro to wirelessly every night using rsync (a built-in command line tool that does incremental backups very quickly, even over a relatively slow 802.11g connection). All it takes is a few tweaks with a tool like Mac HelpMate to get wireless Time Machine backups going instead of Rsync - that's good enough, plus, with the 160gb Firewire disk and 500gb USB disk hooked to the mini, I've got plenty of storage for music and even new video rentals from the iTunes music store, and besides, I'd need its optical drive to install software on the MacBook Air...even if I have to spend $49 for a USB 802.11n adapter for it...

Mac Mini, you're STAYIN'!

The mind can be a very interesting tool when it's presented with a problem and a solution that hits somewhere in between lust and logic, as the MacBook Air did for me. Everywhere on the web I'm reading that the MacBook Air is simply a fancy, expensive, boutique computer that's fit just to be a secondary "adjunct" computer, sort of playing Sunday Corvette to the Minivan of the home office iMac.

But that's not really true...

If there's no need for the heavy-duty processing power of a high-end MacBook Pro, or the 4gb of RAM that high-impact video or high resolution graphic arts dictates, then there's really no need for that iMac. The Mac Mini does the trick, and, with a little love, can double as a Time Capsule device and an AppleTV device.

So let's go over this again...

it's either this, which everyone seems to assume is necessary...

20 inch iMac + Time Capsule + AppleTV + MacBook Air

when in reality, all that necessary (if you're just doing normal SOHO or Consumer stuff) is...

Mac Mini + AirPort Extreme + MacBook Air + 500gb USB + 20 inch display + external hard disk

It's all about the wireless hub now, and the device that sits in the center, and it's still (for now) a Mac computer. Merge the Mini, Airport Extreme, 500gb USB hard drive together, and you've got the AirHub - the future of the consumer (and possibly SOHO) Mac environment.

For everyone who has either considered donating to the OLPC project, or has plunked down the money in 2007 for the give-one-get-one initiative so they could get their hands on a XO for testing or development purposes, I thought it would be nice to share this letter confirming our donation, to affirm that there is indeed a professionally-run system tabulating the contributions.



Some folks might have gotten put off by the fact that the donations originally to go PayPal, and due to the fact that I didn't hear anything for a bit after the initial confirmation, I must confess that I was a little anxious. At any rate, here's the complete letter from the OLPC project confirming our donation. Once again, this really whets my appetite, especially so soon prior to MacWorld where many folks are anticipating a tablet Mac or some other lower-cost solution ripe for educational use.

Dear Donor,

Thank you for participating in the One Laptop per Child "Give One Get One" program.

Your Order Reference Number is: xxxxxxxx26

Your donation of $399.00 will bring education and enlightenment to children of the developing world. $200 of each $399 "Give One Get One" Donation is tax-deductible (your donation minus the fair market value of each laptop you receive). With Shipping and Handling, the total charge to your credit card is $423.95 .

As a "Give One Get One" donor, you will receive one of the first XO laptops to be distributed in North America. Laptops will be delivered on a first come, first served basis.

If you have any questions about your donation, please contact OLPC at service@laptopgiving.org or call 1-800-201-7144. Should your employer wish to match your donation, we are a 501(c)(3) organization and our EIN# is 20-5471780.

T-Mobile USA is proud to offer you one-year complimentary access to T-Mobile HotSpot in recognition of your support to the One Laptop per Child "Give One Get One" program. As you help children in developing countries stay connected, educated and enlightened, T-Mobile is supporting you! You will receive an email message close to the time that your laptop ships containing instructions on how to begin your complimentary year of T-Mobile HotSpot service. Please refer to http://hotspot.t-mobile.com for additional information on T-Mobile HotSpot Wireless broadband Internet service.

To find out more about the mission of One Laptop per Child, please visit www.laptopgiving.org. To review "Give One Get One" terms and conditions, please click here.

Thanks again for your generous support, and welcome to the One Laptop per Child community!

Sincerely,


Nicholas Negroponte

Chairman

OLPC Foundation

One Laptop per Child Foundation is exempt from federal income tax under Section 501(c)(3) of the Internal Revenue Code. Your donation is tax deductible to the extent provided by U.S. law. This receipt is not valid if your donation is refunded for any reason. Additional information about the foundation is available from www.laptop.org or the foundations registration office of your State.

One Laptop per Child

P.O. Box 425087

Cambridge, MA 02142

U.S.A.

We are in the process of closing out the fiscal year here, and it's been a fantastic year for Apple Computer - and MOST Training and Consulting. We feel particularly blessed to be able to work with so many wonderful technologies and so many incredible schools and their support and instructional staff members. Schools can be particularly challenging environments, especially when it comes to showing love to the folks who make all the tech work.

Apple is shattering new sales and profit records, and so are we, on a much, much, much, much (add a million muches) more limited scale. This year, besides sales of Mac HelpMate subscriptions, our greatest area of growth has been in integration and planning projects for K-12 school districts, one-to-one deployments in particular, where each student receives their own MacBook for use at home and at school.

While we of course prefer MacBooks, our good fortune this year led us in search of worthy donations, and we found the OLPC (One Laptop Per Child) initiative too much to resist, especially since we can now donate one and receive one for development purposes. Not only do we have web-based solutions in development that might to be compatible with the OLPC portable, we are also intrigued by the notion of deploying it to students considered too young for their own MacBook - such as kindergarten through third-graders.



I'm also a little concerned that some Apple employees in the Education sector might perceive the OLPC as a threat to the sale of MacBooks, being so much less expensive, but we think they'll be complementary - the "gateway" to the purchase of more expensive Mac OS X computers in later grades.

The OLPC is something we not only have to "see" but something we need to consider as a serious solution for our education customers. We need to find out how it plays on school networks, with content filters, and whether we can manage them using the same tools we count on to manage Mac OS X computers and servers. So, I've sent in my $399 and am anxiously waiting for the green "Martian" with its plastic antennae to arrive.

When Leopard first came out, people started asking the requisite questions: how slow of a Mac can I install Leopard on? Leopard "requires" a G4 processor of at least 867 mhz, and 512 megs of RAM. We have a few older G4s in our lab, one of which has a 350 mhz processor and 1 gig of RAM. I also have an iMac G3 with a 600 mhz processor and 512 megs of RAM.

In order to install Leopard on an unsupported machine, I had a number of options:

1) hack the installer script.

I dismissed this as too time-consuming and possibly illegal, as it might break Apple's EULA

2) put the older Mac into TDM (target disk mode), attach it to a supported Mac with a firewire cable, then boot off of the Leopard Install DVD and install it onto the unsupported Mac that way

That would work, but I need to install/tear-down more frequently, so I went with option #3

3) create a disk image of an install and then use netboot and NetRestore to image an unsupported machine.

The Results:

the G4 350 with a Gig of RAM ran Leopard OK, but with pretty slow screen redraw. It only has a 16 meg ATI Rage Pro for video. I've also subsequently used the same image on a PowerMac G4 Cube 450 mhz, a QuickSilver Tower 800 mhz, with much success, depending on the RAM and strength of video card.

System_Profiler Command Bug

Interestingly (and this is not a show-stopper by any means), the ethernet hardware section of a text-based system_profiler report written to standard output looks like this on two unsupported Macs:

IEEE80211:
Join Mode: Automatic
PreferredNetworks:
SecurityType: WPA Personal
SSID_STR: MOST-N
Unique Network ID: DE79D60B-EE07-417C-ABDC-AE4A5F925435
Unique Password ID: 45721E00-04D8-46C6-8574-AD62648A3E4E
SecurityType: WPA Personal
SSID_STR: AE_Shavit
Unique Network ID: BD1143B7-909F-49B1-B72D-B604CC07026F
Unique Password ID: BB8D64CE-4E60-4E29-93B3-9F57E5CA044B

2008-02-22 12:18:57.169 system_profiler[18657:10b] *** -[RedLightGreenLight unlockWithCondition:]: lock (<RedLightGreenLight: 0x1c0de10> '(null)') unlocked when not locked
2008-02-22 12:18:57.174 system_profiler[18657:10b] *** Break on _NSLockError() to debug.
Modems:

Modem Information:

Modem Model: Spring
Interface Type: Serial
Modulation: V.90
SKU Name: UCJ
Hardware Version: 6.0F
Driver: com.apple.driver.AppleSCCSerial (v1.3.2)
Country:


However,

Any G3 on which I've blasted the image has resulted in a complete failure of hard disk support or a kernel panic. There's not even what I'd call an elegant warning about the processor type being unsupported - and that's not "Apple-like" in my opinion. It's almost as if the G3 is so antiquated that it doesn't deserve a proper send-off. Oh well.

For some inexplicable reason I decided to go with a Sprint for my wireless broadband service after having purchased an iPhone. The Verizon Treo 700p I'd been using served nicely as a broadband dialup device, but the Edge network on the iPhone didn't support tethering (although it's possible with an unsupported hack) and even it it did it wouldn't be enough to use Mac HelpMate in a pinch.

So, after Verizon told me I'd have to pay full price for an EVDO device, I decided to go with Sprint. The fact that I couldn't successfully complete their online registration should have scared me away. Too bad I'm hard to scare.

It took me nearly two hours on the phone to find out how to get the Novatel U727 working - they hadn't shipped it to me with any account information, and after reading the ESN to about the fourth support representative, they were like "Oh" we sent you an unactivated modem, with no associated account information, now, who are you again? What's your Social Security Number?" The inept support made me feel, well, less than confident in this device I was counting on as for emergency internet access.

Once it started working, it was not bad, not bad at all. Until yesterday, when I really needed it in a pinch to demonstrate something to a client, it refused to connect, giving nothing but a "Lock failed" error dialog.



So, after reinstalling the support software three times, and three reboots, the technical support rep asks "Do you have a PC handy?"

My response: "Is this product supported on Mac OS X?"

His reply "Well, yes."

At this point, my suspicions about Sprint's inept tech support are being confirmed, big time. I let the fellow on the phone know that I'm going to try it on another computer. I do, and it works.

At this point, I am too tired to call back Sprint. I decide to have a look at the system log.

And what do I see?

Dec 21 15:05:34 deanmbp pppd[398]: pppd 2.4.2 (Apple version 233-16) started by root, uid 502
Dec 21 15:05:34 deanmbp pppd[398]: Can't create lock file /var/spool/lock/LCK..cu.wwan1d1: No such file or directory

I create the lock directory, and it works again.

I guess once you figure a vendor is incompetent, that incompetence is contagious as well.

Recently, with the price of large (one gigabyte and bigger) keychain (USB) flash drives coming down to below $100 or even $49 when on sale, more and more Mac users and admins are using them as data transfer tool, quickie backup device, or even a data recovery or troubleshooting tool. However, it's easy to forget, in these days of OS X applications that don't have resource forks or create documents with resource forks, that nearly all of these USB drives come formatted with the FAT32 file system, which is still unfriendly to Classic Mac Applications, as well as Postscript Fonts, and QuarkXpress Documents. If you copy file from your Mac to a FAT32 keychain drive, you'll separate the resource forks of the files and find that your Applications, Fonts, and other document may be trashed beyond the point of recoverability.

One way around this problem is to format your USB drive with the HFS+ file system using Disk Utility, however, this prohibits the flash drive from ever being used with a Windows or Linux computer.

I have worked around the problem by using Disk Utility to create a "sparseimage" on my USB keychain drive. A sparseimage is the a disk image with a "growable" filesystem. When you create a sparseimage, you have to specify the maximum size it can grow to, but it won't reach that size unless you fill it up. Now, when you attach your USB drive to a Mac, simply double-click the sparseimage to mount it. Put your Mac files in there, and you'll enjoy the best of both worlds, without having to reformat your USB drive.

Houston, is there a problem?

Recently, on a listserv I frequent, I heard a complaint from a sysadmin that "yet another software RAID mirror" had failed. The fellow complaining emphasized the fact that in the OS X Server boxes he managed, disks would fail and he wouldn't know until he checked in on them. He wondered if any hardware RAID cards featured notification services, rather than rely on the built-in RAID capabilities of OS X.

While an XServe RAID can notify you if a hard disk fails, as well as other hardware RAID subsystems, Disk Utility and OS X have no provisions for notifying you if a mirror set has a bad disk. A mirror is two hard drives acting as one, so that if one fails, you lose no data. You can set up a mirror via the diskutil command or in the Disk Utility application. PCI IDE RAID controllers don't feature notification like external subsystems, as the PCI bus provides no means of communicating the status of various drives or RAID sets. After all, a phone line or ethernet cable is required for the little computer than lives in the RAID chassis, so it can get your attention when it's hurting.

Fortunately, I had a solution at my fingertips.

I had run across this very same problem myself, a couple of years ago, so I hammered out my own little software solution, a shell script that runs as a cron job every hour. It simply checks the status of the RAID to make sure that both disks check out with an "OK" status, take a look at the output of the following command:

host2:~ mostadmin$ diskutil list
/dev/disk0
#: type name size identifier
0: Apple_partition_scheme *34.2 GB disk0
1: Apple_partition_map 31.5 KB disk0s1
2: Apple_Driver43 27.0 KB disk0s2
3: Apple_Driver43 37.0 KB disk0s3
4: Apple_Driver_IOKit 256.0 KB disk0s4
5: Apple_Patches 256.0 KB disk0s5
6: Apple_HFS images 34.2 GB disk0s6
/dev/disk1
#: type name size identifier
0: Apple_partition_scheme *17.0 GB disk1
1: Apple_partition_map 31.5 KB disk1s1
2: Apple_Driver_OpenFirmware 512.0 KB disk1s2
3: Apple_Boot_RAID 17.0 GB disk1s3
/dev/disk2
#: type name size identifier
0: Apple_partition_scheme *17.0 GB disk2
1: Apple_partition_map 31.5 KB disk2s1
2: Apple_Driver_OpenFirmware 512.0 KB disk2s2
3: Apple_Boot_RAID 17.0 GB disk2s3
/dev/disk3
#: type name size identifier
0: boot *17.0 GB disk3

host2:~ mostadmin$

The diskutil list command fetches the identifiers of all the hard disks, in this case, the mirror set we're concerned with is called "boot," which is the boot drive of this OS X Server.

Once you know the identifier of your RAID set, you can use the diskutil command again to check the status, what we're looking for here is "OK." If there's two OKs, everything's hunky-dory. Less than two, you've got a problem. If you've got zero, your server's a paperweight and it won't be spamming you.


host2:~ mostadmin$ diskutil checkRAID disk3
RAID SETS
---------

Name: boot
Unique ID: bootf5b49d82471e11d98b06003065be09be
Type: Mirror
Status: Running
Device Node: disk3
-------------------------------------------------------------
# Device Node Status
-------------------------------------------------------------
0 disk1 OK
1 disk2 OK
-------------------------------------------------------------
host2:~ mostadmin$


Now, for a nice little script I wrote to run as a cron job:


#!/bin/sh
## This script is designed to get the status of a mirror set and send an email notice on fail
## to test, change the "good=2" below to "good=1" and you should receive a warning email
## Dean Shavit, MOST Training & Consulting dean@macworkshops.com http://www.macworkshops.com
## This is for servers using software RAID mirror sets only
##
## Step 1: Define a variable for a functional raid by counting the number of good disks
status=`diskutil checkraid disk3|grep -c OK`
## Step 2: Define a number for comparison against a failed raid
good=2
## Step 3: Define the warning message for the body of the email
warning="houston we have a problem!"
## Step 4: Define a variable for the computer name
box=`/usr/sbin/scutil --get ComputerName`
## Step 5: Define a variable - email address of person to notify
admin="dean@macworkshops.com"
## Step 6: compare current status with good status, if a match, echo, if not, notify
if [ $status == $good ]; then
echo $status
else echo From: $box- $warning!!! > /tmp/houston.txt| mail -s "RAID Alert Report"\
$admin < /tmp/houston.txt
fi

This script will fire off an email if the number of OKs returned by diskutil checkRAID is any number other than 2.

Other Concerns

If you're running Jaguar Server, you'll need to make sure that the mail service is started. Panther or Panther server will use the Postfix-watch process to fire off the email, regardless of whether the email services are running.

Also, make sure that the hostname and/or DNS name of the server is set correctly, and that the email "from" address, which should be root@fqdn (fully qualified domain name) is going out a router or IP address which has a PTR (reverse lookup record) that matches the hostname of the server, otherwise some email servers (maybe yours) might not accept the message.

Here's the email the script generates (this is sent from my Powerbook, which has no RAID nor a proper fqdn):

Received: from ASSP-nospam ([192.168.0.84]) BY 192.168.0.84 ([192.168.0.84])
WITH ESMTP (4D WebSTAR V Mail (5.3.4)); Tue, 01 Feb 2005 01:05:38 -0600
Received: from 68.165.43.42 ([68.165.43.42] helo=minime.local) by ASSP-nospam ; 1 Feb 05 07:05:38 -0000
Received: by minime.local (Postfix, from userid 0)
id D56DB225D6F; Tue, 1 Feb 2005 01:09:47 -0600 (CST)
To: dean@macworkshops.com
Subject: RAID Alert Report
Message-Id: <20050201070947.D56DB225D6F@minime.local>
Date: Tue, 1 Feb 2005 01:09:47 -0600 (CST)
From: root@minime.local (System Administrator)
X-Assp-Spam-Prob: 0.00000
X-UID-FLAGS: r00001665-00000000000000000000000000000000

From: minime- houston we have a problem!!!!

I'm in the process of writing my March column for MacTech Magazine which will feature this solution and many other free solutions for monitoring and repairing hard disks from the command line. If you don't have a subscription yet, get one! Better yet, consider attending one of our workshops where you can learn to roll your own solutions like this script and much, much more. Until then, I hope I don't see you in Houston!

Dean Shavit is an ACSA (Apple Certified System Administrator) who leads training sessions for MOST (Mac OS Training & Consulting). If you have questions or feedback you can contact him at dean@macworkshops.com

Overlapped overlapped extent allocation errors can be the bane of any Mac user's existence. Often, these errors go unnoticed until the problem becomes visible: your Mac might refuse to boot, crash unexpected, or worse, critical data might disappear from the Finder. Disk Utility can detect, but not fix overlapped extent allocation errors, and certain third-party utilities, such as Alsoft Diskwarrior, can fix them, but generally without reporting the consequences.

Overlapped extent allocation error occur when the file system thinks that two files are occupying the same area on the hard disk, hence overlapping on the same "inode," which is the structure which holds the location of the data blocks the file occupies, and also file permissions and flags.

Clearing the "overlapped" or "overallocated" extent allocation essentially means that you'll have to lose some data, because the only way to remove the overlap is to delete the file that's occupying the inode. So, if you suspect, or find out, that the guilty file is a critical system file that resides in one of the hidden system directories such as /etc /var /usr/ or visible system directories such as /System or /Library, and you don't want to reinstall the whole OS (which might not fix the overlapped extent allocation anyway), it's good to have another disk available to copy the files back to your original disk if necessary: a second bootable hard drive or a firewire drive connected to your Mac when you remove the misbehaving file. Just make sure that when you copy the file back to your boot disk that the permissions are correct, so it's best to use the "ditto" command, so that all sticky bits, flags, and permissions are preserved.

In case you didn't know, you don't have to boot from an install CD in order to check for overlapped extent allocations. All you need to do is restart your Mac, while holding down command + S to boot in "single-user mode."

At the command prompt that appears, type:

$ fsck -fy

If you have an overlapped extent allocation, you'll see:

"Overlapped Extent Allocation" (File 123456d)

No matter how many times you run fsck -fy, you'll never be rid of the error.

So, simply issue the following command:

find / -inum 123456 -print

Note the "d" was dropped, or any extra letter that appears after the inode number.

The find will return a file name that machines with the inode number, and the path to that file. If you remove the file then the fsck will not return this error next time you run it.

However, before you can delete the file(s) in single-user mode, you'll need to mount the file system. Type:

$ mount -uw /

When done, issue the "sync" command, and that will flush the write cache so that all pending writes are written from memory to the disk. Also, since most OS X 10.3 Macs use the HFS+ Journaled file system, it might be a good idea to disable the journal before booting into single-user mode by typing:

$ sudo diskutil disableJournal /

then re-enable it when done fixing the overlapped extents and rebooting normally:

$ sudo diskutil enableJournal /

Update: 7/12/2005: Apple has released a Knowledge Base article which reveals an expanded process much like the one we've had posted since last November! Congrats to Apple for finally acknowledging the overlapped extent allocation problem and offering a solution, no matter how geeky! It shows that OS X users are maturing, also with the OS, at least in the opinion of Apple Support!

Chris Anderson is a long-time Linux propellerhead who just got his first Mac, an ibook G4, and can't keep his hands off of it. He currently works as a "The Architect" and general visionary for a maker of world-class collectibles.

Wither SCSI? Maybe it's got more than one life!

It was perhaps two years ago, when OS X 10.2 was all the rage, that I encountered what many Mac Consultants were discovering: certain SCSI adapters and OS X didn't seem to like each other very much. I don't want to go on about it, but anyone who's worked with Macs for the last six years or so has seen SCSI go from a stalwart standard to odd man out. Now, with the advent of the G5 PowerMac with 133 mhz PCI-X slots and 3.3v power requirements, a handful compatible of SCSI host adapters (cards) that once worked with older PCI PowerMacs up through the Mirrored Doors G4s, the last generation towers prior to the G5 has been reduced five possible options, and a few sure bets.

The sure bets for G5 compatible SCSI host adapters are (drum roll):

The Atto UL3D-66 or UL3S-66 for the higher-end G5s with PCI-X Cost: $349/$549
The Atto UL3S or UL3D standard car dor the lower-end G5s. Cost: $349

The Atto ExpressPCI UL4S or ExpressPCI UL4D. Cost: Cost: $349/$549
(UL4 Series cards are really fully compliant with 133mhz PCI-X and deliver whopping 320 mb/sec performance)

Can anyone say ouch? I will! OUCH! That's a hefty price to pay to attach a SCSI tape backup to your XServe G5 or G5 tower, or to support that legacy MO drive you need for your company's digital archives, or that 11 x 17 scanner you can't part with. Someone call a Doctor!

You cannot argue with the quality of Atto's SCSI products, they are the best. However, I can certainly take issue with the cost. The lesson to be learned here is that if you're going to need a SCSI solution for your XServe G5 be sure to purchase the ATTO/Apple OEM build-to-order SCSI solution when you order your server, as it's not available separately.

Old Habits are Hard to Break

I've been accused of being a SCSI nutcase, and I'll admit that I am. I have the following SCSI cards in my collection at the office, some are used, some are not, all bring me a twisted kind of pleasure:

-an Orange Micro SCSI Card (not OS X Compatible, Ultra-Wide, unknown model)
-four Atto UL2D SCSI Cards (sometimes OS X Compatible, not G5 compatible)
-three Atto Express PCI ultrawide cards (OS X Compatible, not G5 compatible)
-two Adaptec 2906 SCSI-2 cards (OS X Compatible, not bootable, not G5 compatible)
-one Adaptec 2930 SCSI-2 card (OS X Compatible, not G5 compatible)
-one Adaptec 2944 Ultra-Wide Differential SCSI card (not Mac Compatible)
-one Adaptec PowerDomain 39160 Ultra-160 card (OS X Compatible, not Retrospect 6 certified, G5 compatible)

It's that last card that nearly made me burst into tears a couple of years ago. A client of mine, a video editor, had purchased a very nice Dual 533 G4 system and serveral large external SCSI drives, along with the vendor's reccomended SCSI host adapter, the Adaptec 39160. It worked OK in OS 9, but occasionally would drop frames when outputting movies to the DV deck. When I upgraded the editor to OS X, the Adaptec card couldn't make it through any session of any length without dropping frames.



The 39160 (pictured above) is one mother of a SCSI adapter. It's got dual channels, two internal 68-pin connectors, and two external VHDCI (very high density SCSI connectors) on the back for external devices. As if that weren't enough, it's also got a 50-pin SCSI-2 connector for internal hard drives. The very design of it screams (R-A-I-D). Unfortunately, it seems like Adaptec let driver support lapse for OS X, leading to Dantz (the makers of Retrospect) to de-certify Adaptec Cards for use with the backup software. Eventually, Adaptec got out an update that attained Dantz re-certification, but by then the damage was done, and, what should have been high-performance 29160 (single channel) and 39160 (dual channel) Ultra-160 cards doing there business in servers and video editing stations found their way onto shelves and Ebay.

And so, the evil 39160 from my client's editing station made its way into my collection, and eventually into a Blue and White G3 server, where it does heavy data lifting every day.

When I got my dual 2 ghz G5, I missed my SCSI peripherals--the personal tape backkup device, my HP Scanjet 4C, my Zip and Jaz drives and two legacy SCSI CD burners. Of course, I know I don't need them, but why throw them away if they're useful. I considered removing the 39160 from the G3 server and putting it in the G5, because I'd heard (surprisingly) that they performed better in the G5 than in any other box. Who knows, maybe the 39160 was engineered ahead of its time?

So, on to Ebay, where I purchased a used 39160 (not specifically Mac, making sure not to purchase the "Dell" OEM cards that were in abundant supply). I got one for $65. Not bad. But when it arrived, it was indeed the Dell card, as pictured below on the model sticker.



So, I decided to flash it with Adaptec's Firmware 1.2, which, incidentally, requires OS 9. Luckily, I have more than a few Macs in our training lab which will boot in OS 9. Upon running the firmware updater, it exclaimed, "This card has no firmware, do you want to update it?" It updated successfully, and now works splendidly in my G5.

A few weeks ago, Adaptec announced that it is going to discontinue support for all of its Mac SCSI cards, and it's a toss up if the last driver for the 39160 card released this January will work with Tiger or Tiger server. Hopefully, it will, as I'm finding it to be a rock solid, fast, and cheap fix for my SCSI addiction.

Dean Shavit is an ACSA (Apple Certified System Administrator) who leads training sessions for MOST (Mac OS Training & Consulting). If you have questions or feedback you can contact him at dean@macworkshops.com

So if you're like me... you love VNC - a brilliant, open source, remote display technology, that's platform agnostic.
Take control your OS X Mac from a PC, Linux, Palm or just about anything.
Thankfully the server has been available for OS X for quite some time courtesy of OSXvnc

Often I find myself in a situation where I'd like to use VNC on a machine that I only have SSH access to. Problem you say?
Hardly... just a few simple command and you'll be up and running!

1. SSH to the machine you want desktop/aqua access to
2. Download VNC
   1. curl -O http://aleron.dl.sourceforge.net/sourceforge/osxvnc/OSXvnc1.33.dmg
3. Mount the DMG and copy the files to the Utilities folder
   1. hdiutil mount OSXvnc1.33.dmg
   2. cd /Volumes/OSXvnc
   3. cp -R OSXvnc.app /Applications/Utilities
   4. cd /Applications/Utilities/OSXvnc.app
4. We need to setup a password file before launching VNC, so
   1. ./storepasswd mysecurepassword .osxvncauth
5. That's pretty much it... we're ready to go. Launch the server with
   1. ./OSXvnc-server -rfbport 5900 -desktop mymac.local -dontdisconnect -restartonuserswitch N -keyboardLoading N -pressModsForKeys N -swapButtons -rfbauth /Applications/Utilities/OSXvnc.app/.osxvncauth &

We now can connect to the IP of the machine using the password of "mysecurepassword" Just fire up your favorite VNC client on whatever operating system you choose, and connect to the server. Since we used the '&' to load the process in the background, use ps to grab the PID and kill the server when you're done. Also, I've assumed that you aren't running a firewall blocking the vnc port - if you can't connect you might want to check that out.

There's lots of other great options, like tunneling the VNC server through your SSH connection and setting up VNC as a startup item from the command line. Perhaps I can cover that in another article if anyone's interested.

As for me -- I'm Adam Lavelle -- Consultant and Trainer for MOST. I can be contacted at Adam[at]MostConsulting.com