- Combo Updater Differences
- Username bug in Windows environments; workaround
- Booting Issues; what is in PRAM
- The much debated "memory leak"; Using Cocktail
- Wake from Deep Sleep issues: Kernel Panics, fix; Sonnet; Cinema Displays
- Strange screen saver problems
- Shared Libraries issues
- Problems with FrameMaker in Classic
- Problem with multiple displays
- GPRS flaw fixed
- iPod recognition; fix
Mac OS X 10.2.6 represents not only one of the quickest release turnovers from Apple, but also one of the fastest responses by virtually any OS developer to a major flaw - in this case repeated kernel panics caused by a bug in the USB architecture. The list of enhancements, according to Apple's documentation, is as follows:
- Addresses an issue for Mac OS X 10.2.5 in which a kernel panic message may appear if certain USB hubs or devices are connected.
- Improves printing compatibility for PostScript OpenType fonts.
- Improves compatibility for Maya Auto Paint and Paint Effects features.
- Addresses an issue for Mac OS X 10.2.5 in which English is the default language and Asian language scripts are not be available in the International preference pane, if a script's font is not available.
- Improves compatibility for MacSoft Unreal Tournament 2003 on computers with GeForce 2MX and GeForce4MX graphic accelerator chipsets.
- Address Book now supports the Sony Ericsson T610 phone.
Combo Updater Differences
As we have explained in numerous past Mac OS X troubleshooting reports, applying the combo updater rather than the incremental updater for Mac OS X 10.2.6 and other releases can alleviate a broad range of problems. MacFixIt reader Philip Tejada set up a study of sorts to see exactly what differences there are between identical systems: one updated with the incremental updater, and the other with the combo updater. The results are astonishing:
"I set up two identical partitions with an identical copy of 10.2.5 on each. One I updated with the incremental updater, and the other with the Combo updater. What I found is that over 2800 files were different on the two systems after the updates. Many of these files are the actual binaries of applications and frameworks.
"I would have assumed that running either updater should yield the same system, but that assumption proved to be woefully wrong. I've included a text file with the different output comparing the two systems. With the exception of just a few files that I would expect to be different after booting both systems, there are many files that obviously shouldn't be different."
Obviously, the lack of presence, or "staleness" of these files can be the root of a host of issues. Tracking down the individual culprits can be difficult, so applying the usually large (80 MB ) combo updater is your best bet.
Username bug in Windows environments; workaround
MacFixIt reader Brian Bard describes an odd bug in Mac OS X's Windows networking infrastructure that does not allows users to have login names that begin with the character "x."
"In order to fit our new Mac OS X clients into our enterprise network, the OS X short name must match the Windows short name on the password server. Company standards dictate how short names are assigned. Typically, the first initial of the first name is followed by the first seven characters of the last name (i.e., Jimmy Jonestown would be jjonesto). But, for some users (i.e., temps and consultants), an 'x' is followed by the first initial of the first name which is then followed by the first six characters of the last name (i.e., Jimmy Jonestown would be xjjonest).
Bard's problem is that if the first letter of the short name is "x" and there is more than one user (i.e., ouradmin and xjjonest) the short name text becomes garbled when viewed in the "Ownership and Permissions" pane of the Finder "Get Info" window. He writes:
"Sometimes the short name is displayed as Control and Option characters followed by '(Me)' and sometimes it is displayed as a blank line. In addition, the 'www' user may also turn into a blank line. It all depends on how many users there are and how many have 'x' or 'y' as the first character of their short names. I've tested this with 'x' and 'y' but I suspect any character after (but not including) "w" will produce this behavior.
"Everything works. You just have to guess at to whom you are assigning ownership when using the Finder's 'Get Info' window. And, FileXaminer correctly reflects your Finder choices in perfectly readable text. This simply appears to be a Finder display bug."
After some further investigation, it appears that this bug occurs with every letter after "w." Some users have had success with the utility NIM Repair, which can make it easier to rebuild and save changes to the NetInfo Manager - where various login and password information is contained.
Booting Issues; what is in PRAM
Mac OS X Install Disc will not boot Several readers in the MacFixIt Forums - all (perhaps unrelated) using flat-screen iMacs - are reporting problems booting from the Mac OS X Install Disc since upgrading to Mac OS X 10.2.6 .
Reader Robert Shalit writes:
"I have tried every possible restart/start in the book: selecting the CD in System Preferences/Startup Disk, restart with Option key, restart with C key, restart from the Welcome to Mac OS X window/Install icon on the CD, etc." Zapping the PRAM (by holding down Command-Option-P-R at startup) has been reported to work in some cases, but certainly does appear a consistent solution.
Speaking of PRAM, the MacFixIt staff came across this dated document on Adobe's support site about what exactly is contained in PRAM:
- validity status of the clock chip
- clock settings (alarm setting, time, date)
- node ID hint for a modem port (9600 baud, 8 data bits, 2 stop bits, no parity) and printer port
- modem port configuration stop bits
- printer port configuration and connection
- printer connection
- monitor colors
- application font number minus 1
- auto-key threshold and rate
- speaker volume
- double-click time
- caret-blink time
- menu blink
- mouse scaling
- preferred system start-up disk
- network setting (network defaults to LocalTalk)
Technically, if one of these status indicators is corrupt or damaged, it can prevent proper Macintosh startup or cause other problems. Again, however, zapping the PRAM has never emerged as a genuine, across-the-board solution for any one troubleshooting issue.
Cold-boot from external drives can fail; workaround Mac OS X 10.2.6 can cold-boot correctly from external FireWire disks (in our case, a WiebeTech 20 GB) using some Macs, but never using other Macs, such as the iMac G3/700.
In older iMac (and others systems') case, the Mac always boots from the internal disk (IDE/ATA) from a cold-boot (starting up the machine from an off position), but will boot from the external drive if you restart.
Pressing the Option key while cold-booting does not show the external FireWire disk to choose it; not even rescanning, or unplugging-plugging the FireWire cable at that stage and then rescanning will display the disk.
Note that during all these tests, the external FireWire disk is always selected as the preferred booting device in System Preferences.
Some CRT iMac users have had success with the keyboard combination "Command-Option-Shift-Delete". This should bypass the primary startup volume and seek a CD or external drive.
The much debated "memory leak"; Using Cocktail
Before we delve into the Mac OS X 10.2.6 "memory leak" issue, it is important to note that excessive pageouts - the most corroborated symptom after applying the update - do not necessarily cause slower system responsiveness.
Pageouts only occur (in general) to inactive applications and therefore have little or no effect on the active one. In fact, a sophisticated OS can produce better performance by intelligent use of swapping and paging (in and out). To put it another way, it can be quicker to "swap in" a "swapped out" application (or part, thereof) than it is to launch it. As one MacFixIt reader writes:
"We need to remember that paging and swapping have been a feature of many Operating Systems since the late 60's - it is not as though it is a new technology."
That being said, it is undeniable that hundreds of readers - and countless others who are not even aware of the problem - are experiencing a significant overall system slow-down that appear to be memory related. Some users link the problem to individual applications; several corroborated reports indicate a spike in memory occupation when iPhoto is launched.
MacFixIt reader Luke Hertenstein writes:
"What bothers me is that my VM (virtual memory) pageouts are through the roof. After less than 24 hours of up-time, using only a handful of office and Internet apps (no graphics or design apps), I have over 8000 pageouts. Prior to Mac OS X 10.2.6 , I could have a week of uptime that included some design work and still have zero pageouts.
"Something has definitely changed, and it is very annoying. After about a half day of usage, everything seems to slow down and I get lots of 2-5 second pauses (similar in feel to the audio pause on laptops, but it happens ever ten seconds rather than every 30). I am having to reboot at least once a day to clear up my RAM."
Run Cocktail The all-in-one system repair and maintenance utility Cocktail (which accesses cron scripts and other clean-up functions in Mac OS X) has dissolved memory sprawl issues for a number of readers. Some users who were previously seeing Safari and iPhoto chew up several hundred MB of RAM report that the iApps now stick to a reasonable chunk of memory.
Wake from Deep Sleep issues: Kernel Panics, fix; Sonnet; Cinema Displays
Kernel Panics, fix If you are having problems with kernel panics every time you wake your Mac from deep sleep, try deleting the file "com.apple.Powermanagement.xml" located in "Macintosh/HD/private/var/db/SystemConfiguration". After restart you should be able to wake from deep sleep without incident.
Sonnet ATA133 Several readers are experiencing a wake-from-deep-sleep issue with Sonnet's ATA133 version 4.0 drivers with the Sonnet Tempo Trio ATA133 PCI card.
Sonnet's Firmware v3.2.5 with the PCI Tempo Trio card had a known issue waking from deep sleep in Mac OS X 10.2.3 and earlier. Mac OS X 10.2.4, 10.2.5, and 10.2.6 fixed this problem.
However, users report that after upgrading the Sonnet Tempo Trio ATA 133 PCI card with the "tempo133_firm_40" firmware updater from v3.2.5 to v4.0, under Mac OS X 10.2.6 , the inability to wake from deep sleep re-occurs. A hard restart is required in order to regain system functionality,
Downgrading to the Sonnet 3.2.5 firmware will again allow normal sleep.
Cinema Displays users with Cinema displays are having problems making their display enter standby mode.
The problem only occurs when the system is put to sleep automatically after the interval specified in System Preferences. In most cases, if you manually put your system to sleep, the display will enter standby mode properly.
Strange screen saver problems
It seems that after upgrading to Mac OS X 10.2.6, some users are unable to activate screensavers via the upper left hand corner when using "hot corners." Strangely, switching to other corners (and de-selecting the upper left hand corner) will allow screen saver activation for this handful of affected users.
Shared Libraries Problems
If you are having problems with applications freezing at launch after upgrading to Mac OS X 10.2.6 , trying using the Apple Disk Utility to repair permissions (good practice after any incremental update).For some users, the problem only goes away temporarily after repairing permissions on the following files:
Problem with multiple displays
Since upgrading to Mac OS X 10.2.6 some users are noticing that every time the computer wakes from sleep one monitor is unrecognized (it remains in power-down mode) and the other switches to a low screen resolution. One MacFixIt reader writes:
"If I open the 'Displays' preferences panel and hit the 'Detect Displays' button the inactive monitor turns on for a second, but then goes back to sleep. Log out and back in doesn't help. I have to restart the computer to get the 2nd display to work."
Problems with FrameMaker in Classic
Martin Ley describes a repeatable (and in-house confirmed) problem that occurs in the Classic environment after upgrading to Mac OS X 10.2.6 :
"Adobe FrameMaker 7 (a classic app) now displays empty space at the beginning and end of files as white space, instead of grey. For example, if you're scrolling using 'Facing Pages' and the first page is a right-hand page, you'll see a totally white space on the left."
GPRS flaw fixed
Jan Fuellemann at Nova Media sent us a note regarding Mac OS X 10.2.6 's resolution of a GPRS modem bug:
"We recently posted news about non-working GPRS PC Card modems for the new 17" PowerBook due to lack of power supply. Nova media is happy to report that this issue was solved by Apple with the Mac OS X 10.2.6 update."
iPod recognitionWe have received corroborating reports indicating loss of iPod recognition after upgrading to Mac OS X 10.2.6 . This quirk springs up with every refresh to Mac OS X, and is usually solved by applying the combo updater.Resources