[Jim Stoneburner]

This has all the earmarks of a memory leak....

I have been plagued with slowing performance over time, requiring
reboot about once per day of my Pismo G3/500. Application launches
begin to take twice the normal time and more.

This appeared after installation of 10.2.6.

Symptoms

First, one notices that launched apps don't seem to display their
windows as fast as normal. The distinctive chatter of hard disk pageouts
is a clear sign. As more time passes, the application name itself is slow
to appear in the menubar upon launch. Eventually, merely switching
between apps is met with long delays.

My System

Before installing the update over a working 10.2.5 system, I ran DiskWarrior 2.1.1 and repaired permissions.

The following were done after installing 10.2.6 on my Pismo G3/500:
- Repaired and rebuilt the disk directory using DiskWarrior 3.0
- Repaired permissions
- Optimized the hard disk using Speed Disk and the freeware OS X optimization file.

I wasn't sure of the cause, because I run a few add-on programs (all the
latest versions) which I speculated might conflict with the new 10.2.6 update:

Fruitmenu
APE Manager
Diablotin
PowerChute
TinkerTool
Hotsync's Transport Monitor
PTHPasteboard
PTHClock
Escapepod
TypeIt4Me
External devices (Firewire, card readers, UPS, keyboard, others)

I used the combined updater to update from 10.2.4.

Has anyone confirmed this memory leak on a fresh installation?
No applications have been updated, except for Safari. So the memory leak is either in Safari or in OS X itself, at least the parts that run on my computer.

I skipped 10.2.5 due to USB problem warnings.
I had to modify 10.2.4 on our iMac to get the dialup to work.
Now 10.2.6 has a crippling memory leak?
When will I learn not to install updates so quickly?
When will Apple learn to test their updates more thoroughly?

--

JIM STONEBURNER
President, Sonolithics

[Esquare]

Originally posted by Jim Stoneburner:
This has all the earmarks of a memory leak....

I have been plagued with slowing performance over time, requiring
reboot about once per day of my Pismo G3/500. Application launches
begin to take twice the normal time and more.

[...]

When will I learn not to install updates so quickly?
When will Apple learn to test their updates more thoroughly?

When will you learn to do some troubleshooting before you start ranting?

A good way to figure out if something is wrong with your system or with one of the zillion haxies you installed, is creating a new user (call it Test User). This one will have a fresh set of preferences and a 'clean' OS. If the problem disappears when you're logged in as Test User, which seems very likely to me, at least you know where you have to start looking.

EE

[Jim Stoneburner]

You may not have noticed that this memory leak (in the OS or Safari perhaps) was reported elsewhere. Taken in that context, a report like mine which falls short of thorough testing is useful. If it isn't to you, you needn't post. ;^)

While I agree that doing what you suggest is a more sure fire way to prove the hypothesis (wouldn't a reinstall be even moreso?), it isn't very useful when one has to do work on the computer on an ongoing basis. A problem or symptom can be reported, and its cause speculated upon, without proof.

Besides, if you check elsewhere on the web (or the beta testing I've done privately, and troubleshooting of computer-based systems for 20 years), you'll find that my shooting from the hip is usually pretty accurate.

A thesaurus provides us with lots of words that fall short of proof: hypothesis, speculation, supposition, idea, feedback... All of these are useful. You just have to know how to tell the difference between them and proof.

Robert Bly told the story that his late mentor wrote a poem each morning after rising. Robert inquired how to deal with the fact that most of them aren't going to be any good. His mentor said "I lower my standards." ;^)

[Gee4orce]

Well, here are several constructive things you can do to isolate the problem:
[*] show us the top part of the output of 'top'[*] show us a listing of your swapfiles (ls -l /private/var/vm)[*] quit Safari. Do things improve ? Relaunch.

I've been running 10.2.6 since it came out, and have been hammering my system pretty hard. My uptime is 13 days, and I've got 10 swapfiles (which seems like a lot), but haven't seen any performance problems at all.

Just because you experience a problem on your machine it's not justification for criticising Apple's QC - in my experience it's well above average for the software industry.

[Esquare]

Originally posted by Jim Stoneburner:
A load of semantics

OK. My hypothesis is that it isn't a memory leak in the OS or Safari as such, because of the simple fact that neither of my Macs shows any signs of it. Ergo, my speculation is that this is something specific for your setup.

I can't go any lower than that!

EE

[OpenStep]

It's probably Safari. It started running slow for me a couple days ago and when I quit it started grinding eating up swap space. I restarted and got my space back. Try using IE for a couple days and see if it solves your swap problems.

[hudson1]

Originally posted by OpenStep:
It's probably Safari. It started running slow for me a couple days ago and when I quit it started grinding eating up swap space. I restarted and got my space back. Try using IE for a couple days and see if it solves your swap problems.

I noticed this, too, a few days ago. Safari came to almost dead stop. I don't recall if I just logged out or did a restart to fix the symptom but the problem hasn't occured since then. I didn't check what was going on with swap files but it wouldn't surprise me if Safari was spawning a number of them.

[kbata]

Safari has an option to empty it's cache. There are other utilities that will clear out cache's system wide. When the system is running slow try running process viewer and look to see if something is eating up the processor. Since you suspect a memory problem you might try running an application that monitors memory usage.
One of the biggest problems with the web is that someone jumps to a wrong conclusion about a problem, writes a story about it and people start thinking it's fact. It's impossible for Apple to test eery hardware combination and software combination, especially systems that have ton's of hacks installed.

[wulf]

I was curious to see if anyone on MacNN was having this problem, after seeing the article about it on MacFixIt. But I'm not really arguing the point either way, since I haven't experienced these symptoms either (running two machines continuously since updating to 10.2.6).

Many posters (on that thread) omit such pertinant info as: amount of RAM, how many apps are running, and even how much disk space remains on / , all of which could help in tracking down the problem...

Has anyone else here experienced this "memory leak"? Please post (with specs) if you have, and it might go some way towards helping others...

[suthercd]

No memory leaks in Safari 1.0 Beta 2 (v74) or in 10.2.6 than I can detect. The suggestion to Empty Cache has helped me in the past as well as looking at the cookies retained and deleting those that do not serve my needs. I had a problem after upgrading/replacing Safari 1.0 (v72) with web page updates, weird graphics, font handling and other display irregularities that the cookie editng solved.

Also try reboot into single user mode:
Command-s after the chime:
Run /sbin/fsck -y to do a filesytem consistency check.

The -y parameter supplies a default yes answer to all questions asked by fsck.

Continue to run /sbin/fsck -y until no errors are reported. Enter Reboot.

HTH
Craig

[eno]

Originally posted by Gee4orce:
[*] show us a listing of your swapfiles (ls -l /private/var/vm)

And how do you interpret those results?

For example, what do you make of this?

drwx--x--x 12 root wheel 408 May 10 11:16 app_profile
-rw------T 1 root wheel 80000000 May 18 10:07 swapfile0
-rw------T 1 root wheel 80000000 May 18 10:28 swapfile1
-rw------T 1 root wheel 80000000 May 19 03:41 swapfile10
-rw------T 1 root wheel 80000000 May 21 11:34 swapfile11
-rw------T 1 root wheel 80000000 May 21 13:04 swapfile12
-rw------T 1 root wheel 80000000 May 21 13:41 swapfile13
-rw------T 1 root wheel 80000000 May 18 10:29 swapfile2
-rw------T 1 root wheel 80000000 May 18 10:29 swapfile3
-rw------T 1 root wheel 80000000 May 18 10:31 swapfile4
-rw------T 1 root wheel 80000000 May 18 10:43 swapfile5
-rw------T 1 root wheel 80000000 May 18 10:49 swapfile6
-rw------T 1 root wheel 80000000 May 18 11:01 swapfile7
-rw------T 1 root wheel 80000000 May 19 03:40 swapfile8
-rw------T 1 root wheel 80000000 May 19 03:40 swapfile9

uptime:
12:23AM up 3 days, 14:16, 4 users, load averages: 0.33, 0.40, 0.59

output from top of "top":

Processes: 71 total, 2 running, 69 sleeping... 194 threads 00:23:26
Load Avg: 0.63, 0.45, 0.31 CPU usage: 14.3% user, 12.3% sys, 73.4% idle
SharedLibs: num = 123, resident = 24.0M code, 2.45M data, 7.87M LinkEdit
MemRegions: num = 14617, resident = 396M + 11.8M private, 124M shared
PhysMem: 97.0M wired, 427M active, 447M inactive, 971M used, 52.8M free
VM: 5.50G + 75.6M 396874(0) pageins, 1361075(0) pageouts

[normyzo]

eno, can you give us a ps -aux too?

that does seem like a lot of swap files and a lot of memory used, but if you have a lot of programs/processes open that would explain it. If, on the other hand, you have nothing open, there's a problem.

Dan

[crazysprocket]

I just noticed I have a lot of pageouts also, on a machine that hasn't historically had very many (on 10.2.4/5) with the same workload it has had all along.

Processes: 54 total, 3 running, 51 sleeping... 136 threads 12:40:25
Load Avg: 0.17, 0.22, 0.28 CPU usage: 0.9% user, 12.9% sys, 86.2% idle
SharedLibs: num = 72, resident = 23.9M code, 1.93M data, 7.42M LinkEdit
MemRegions: num = 5621, resident = 141M + 8.93M private, 100M shared
PhysMem: 63.4M wired, 247M active, 158M inactive, 468M used, 44.0M free
VM: 2.04G + 42.2M 43720(0) pageins, 27751(0) pageouts

[soja:~] jeff% ls -l /var/vm/
total 468768
-rw------T 1 root wheel 80000000 Apr 28 18:45 swapfile0
-rw------T 1 root wheel 80000000 May 9 18:59 swapfile1
-rw------T 1 root wheel 80000000 May 12 01:10 swapfile2


Also 2 more swapfiles than I'm used to having.

I kind of think there's a new leak in 10.2.6 (or maybe Safari) that wasn't there before.

[suthercd]

Discussion of 10.2.6 memory leaks is under way at Macfixit

Craig

[IamBob]

eno, can you give us a ps -aux too?

That should be `ps -auxc` or `ps -auxww`.

Otherwise you can't really tell what the processes are. With the 'c', you get just the app names (minus path and arguments). With the 'ww' and not the 'c', you get the path and arguments in their entirety. With neither the 'c' or 'ww', you get the path and arguments clipped...resulting in (mostly) un-readable output.

Anyway, I've experienced the same problem. I have exactly one process that munches the ram - System Events. I've tested multiple versions; 1.1.1 and 1.2. Both have the same problem.

I'll test a new user later and post-back.

[Jim Stoneburner]

I've rebooted recently, so no thrashing or delay is seen right now. When it reappears, I'll try these until we can identify what's up:

In Terminal:
top, or top -u (look at CPU usage, etc)
ls -l /private/var/vm (look at virtual memory usage)
ps -auxc (memory usage)

Other steps:
* Quit apps one at a time to see if problem abates.
* Turn off third-party stuff and try whole process again.
* Boot into new user and try process again.

What have I missed?

Already done:
* No need to do another fsck -y, because I ran DiskWarrior a couple of times (its more complete than fsck)
* updated prebinding
* repaired permissions
* I've also optimized disk space recently.
* The 10.2.6 update was the combo updater over 10.2.4 after a fresh DW run and permissions repaired.

PS: I have 640MB physical ram, and over 4GB hard disk space.

I've never had a problem that a required clean install or archive and
install, in over 18 years of Mac usage. It always turns out to be directory
or wrapper damage, a buggy or incompatible app or other add-on, an
OS bug (yes there are many), or a particular corrupt preference.
Reinstalls are recommended too early, I feel.

I know it may seem unlikely for the OS to have a memory
leak, and it may indeed turn out to be an app or add-on. But, keep in
mind that most OS bugs - the recent USB bug for instance, or the
previous dialup disconnection issue (for which I was the author of the
original AppleCare thread) - only show up under certain circumstances
and configurations. They were real, they were fixed by Apple, and they
did NOT require (or benefit from) clean installs and the like.
(10.2.4 software modem disconnect issue)

[mrmister]

I, on the other hand, have a bunch of haxies installed and am experiencing no slow downs and no pageouts on 10.2.6...for what it's worth.

I don't use Safari, though.

[eno]

Originally posted by normyzo:
eno, can you give us a ps -aux too?

Here is the output of ps -auxcm (the "m" switch sorts by memory use):

Code:
USER      PID %CPU %MEM      VSZ    RSS  TT  STAT STARTED      TIME COMMAND
enofile 18279   0.0 15.5   786256 162976  ??  S     5:27AM   0:53.18 Proteus
enofile 17688   0.0 11.3   929000 118412  ??  S    12:24PM  16:52.83 Safari
enofile 18276   0.0  7.5   794712  78972  ??  S     5:27AM  12:38.29 Mail
enofile   179   0.0  7.0   821780  73544  ??  Ss   Sun10AM 267:21.82 Window Manager
enofile   388   0.0  2.6   798984  27708  ??  S    Sun10AM   9:25.04 Finder
enofile   424   0.0  2.0   681048  20996  ??  S    Sun10AM   4:20.97 LaunchBar
enofile 15261   6.0  1.9   773940  19740  ??  S    Tue05PM  27:31.73 Project Builder
enofile 10822   1.8  1.7   675300  17652  ??  S    Mon07PM 224:56.10 SystemUIServer
enofile 17446   0.0  1.6   653824  16980  ??  S     7:43AM   1:46.35 Terminal
enofile 18284   0.0  1.5   643740  15892  ??  S     5:27AM   0:01.45 LaunchCFMApp
enofile 17691   0.0  1.0   709284  10204  ??  S    12:29PM  12:37.32 LaunchCFMApp
enofile 17515   0.0  0.9   640284   9144  ??  S     7:51AM   0:06.11 JellyfiSSH 2.1.2
enofile   569   0.0  0.8   695792   8696  ??  S    Sun10AM  82:29.26 Dock
enofile   434   0.1  0.7   646116   7220  ??  S    Sun10AM  35:16.36 Net Monitor
enofile   421   0.0  0.6   651532   6600  ??  S    Sun10AM   0:40.40 LiteSwitch X
enofile   364   0.0  0.5   115360   5744  ??  Ss   Sun10AM   2:25.11 loginwindow
enofile   441   0.0  0.4   647444   4476  ??  S    Sun10AM  12:46.66 Default Folder X
enofile   435   0.0  0.4   644292   4152  ??  S    Sun10AM  13:02.97 LaunchCFMApp
enofile   650   0.0  0.3   638924   3592  ??  S    Sun10AM   0:06.32 SecurityAgent
root      337   0.0  0.3    66464   3192  ??  Ss   Sun10AM   0:27.36 coreservicesd
enofile   475   0.0  0.3   635016   2916  ??  S    Sun10AM   9:25.15 ASM_Controller
enofile   175   0.0  0.3    59280   2648  ??  Ss   Sun10AM  16:56.79 ATSServer
enofile   429   0.0  0.2   637232   2184  ??  S    Sun10AM   8:08.56 UniversalAccess
enofile 18283   0.0  0.2    21952   2032  ??  S     5:27AM   0:00.20 imservices
enofile 18286   0.0  0.2    17300   1828  ??  Ss    5:27AM   0:06.20 imserviced
enofile   438   0.0  0.1   633744   1416  ??  S    Sun10AM   0:17.32 escapepod
enofile   428   0.0  0.1   633596   1404  ??  S    Sun10AM   1:21.80 USBOverdriveHelp
enofile   594   0.0  0.1    17776   1312  ??  S    Sun10AM   0:07.16 AppleSpell
root      372   0.0  0.1    21868   1160  ??  S    Sun10AM   0:02.64 DirectoryService
enofile 18385   0.0  0.1     6164   1160  p2  Ss    6:41AM   0:00.14 tcsh
enofile 18342   0.0  0.1     6164   1144 std  S     6:33AM   0:00.21 tcsh
enofile 18287   0.0  0.1    14732   1096  ??  Ss    5:27AM   0:04.48 imserviced
enofile 18285   0.0  0.1    14576   1028  ??  Ss    5:27AM   0:05.71 imserviced
root    18138   0.0  0.1    15220    904  ??  Ss    2:48AM   0:00.53 lookupd
root      116   0.0  0.1    16344    776  ??  Ss   Sun10AM   3:10.26 configd
root      142   0.0  0.1    15496    728  ??  Ss   Sun10AM   0:03.81 SecurityServer
enofile 18398   0.0  0.1     1792    632  p2  S+    6:41AM   0:00.03 ssh
root    18341   0.0  0.1    14048    568 std  Ss    6:33AM   0:00.30 login
root      407   0.0  0.0    16208    436  ??  Ss   Sun10AM   0:24.26 cupsd
enofile   379   0.0  0.0    16028    416  ??  Ss   Sun10AM   0:05.20 pbs
root    18070   0.0  0.0    14464    348  ??  Ss    2:21AM   0:00.37 pppd
root    18419   0.0  0.0     1376    332 std  R+    6:45AM   0:00.00 ps
root      310   0.0  0.0     1700    276  ??  Ss   Sun10AM   0:08.88 netinfod
root      206   0.0  0.0    14992    276  ??  Ss   Sun10AM   0:07.47 autodiskmount
enofile   386   0.0  0.0    14312    248  ??  S    Sun10AM   0:02.85 aped
enofile 18420   0.0  0.0     1060    204 std  R+    6:45AM   0:00.00 sed
root      288   0.0  0.0    14516    188  ??  Ss   Sun10AM   0:00.57 mDNSResponder
root        2   0.0  0.0     1844    176  ??  Ss   Sun10AM   0:07.70 mach_init
root      356   0.0  0.0    15008    160  ??  Ss   Sun10AM   0:00.15 automount
root      390   0.0  0.0     1556    124  ??  Ss   Sun10AM   0:26.30 ntpd
root      290   0.0  0.0    13584    108  ??  Ss   Sun10AM   0:00.08 crashreporterd
root      406   0.0  0.0    13868    104  ??  Ss   Sun10AM   0:01.01 cron
root      279   0.0  0.0     1308    100  ??  Ss   Sun10AM   0:02.02 syslogd
root        1   0.0  0.0     1308     48  ??  Ss   Sun10AM   0:00.05 init
root       81   0.0  0.0     1292     36  ??  Ss   Sun10AM   1:40.35 update
root       85   0.0  0.0     1316     24  ??  Ss   Sun10AM   2:02.51 dynamic_pager
root       51   0.0  0.0    15272      8  ??  Ss   Sun10AM   0:01.44 kextd
root      404   0.0  0.0     1308      8  ??  Ss   Sun10AM   0:00.00 inetd
root      348   0.0  0.0     1296      8  ??  S    Sun10AM   0:00.00 nfsiod
root      347   0.0  0.0     1296      8  ??  S    Sun10AM   0:00.00 nfsiod
root      346   0.0  0.0     1296      8  ??  S    Sun10AM   0:00.00 nfsiod
root      345   0.0  0.0     1296      8  ??  S    Sun10AM   0:00.01 nfsiod
root     5175   0.0  0.0        0      0 con- Z     1Jan70   0:00.00 Mail
root    17416   0.0  0.0        0      0 con- Z     1Jan70   0:00.00 Mail
root    16093   0.0  0.0        0      0 con- Z     1Jan70   0:00.00 Mail
root    16027   0.0  0.0        0      0 con- Z     1Jan70   0:00.00 Mail
root    14192   0.0  0.0        0      0 con- Z     1Jan70   0:00.00 Mail
root    13021   0.0  0.0        0      0 con- Z     1Jan70   0:00.00 Mail
root    11377   0.0  0.0        0      0 con- Z     1Jan70   0:00.00 Mail
root    10471   0.0  0.0        0      0 con- Z     1Jan70   0:00.00 Mail
root     7270   0.0  0.0        0      0 con- Z     1Jan70   0:00.00 Mail
root      647   0.0  0.0        0      0 con- Z     1Jan70   0:00.00 Mail

And here is the output of vm_stat:

Code:
Mach Virtual Memory Statistics: (page size of 4096 bytes)
Pages free:                     4157.
Pages active:                 153578.
Pages inactive:                80067.
Pages wired down:              24342.
"Translation faults":      262116572.
Pages copy-on-write:          832371.
Pages zero filled:         131673191.
Pages reactivated:           5766490.
Pageins:                      405982.
Pageouts:                    1370997.
Object cache: 380832 hits of 584743 lookups (65% hit rate)

[suthercd]

Jim-


Been thinking and looking for ideas, but none yet.

I run 10.2.6 on 4 machines- 3 G4s and one iBook G3 and use Safari as a primary browser. There are two situations where I have seen this kind of dramatic performance degradation. One is on the iBook (512K) when iTunes 3, Limewire, and IE were running. IE and iTunes 2 were infamous memory hogs, but quitting one or all would relieve the slow app switching, window zooming, and general sense everything delayed. The second was in some of the developer releases of 10.2.5, but these were very buggy- Mail and Finder particularly. That said, all machines operate as expected today with 10.2.6.

The problems appear to be particular to your environment- or so I thought until reading the MacFixit threads on this. So now, it looks like there are a percenatge of machines that exhibit this slow down. What's common about them? If it were a software/system problem, the complaints would be more widespread. (?)

Curious to see what happens when you create a new user and log into that account. Hopefully all will work fine and so the problem lies in prefpanes/apps. I have not yet found that disabling haxies and app support 3rd party stuff helps solve OS X problems- very different fom 9 back experiences. I have found that having an out of date sahreware app can cause window manager problems as well as just plain weird behavior. I'm sure you have checked into that.

If your still having the problem, could your run vm and ps with the above parameters and post the results.? If that's too much, I understand.

Craig

[ret]

When I saw the topic heading, I thought "Can this solve my problem?" But apparently not. My dramas with 10.2.6 are a little different.

Since installing the update my Ti500 has slown to be virtually unusable. However, this manifests itself in a start-up time of around 30 minutes! Once the machine is running, Memory Monitor shows paging occurring for virtually every action (switching applications, windows, menus etc) I get similar problems to other posters here with endlessly bouncing apps starting up and lots of HD activity.

Here's the kicker tho, even booting into my Classic partition exhibits similar behaviour. This makes me wonder if something in the lower level disk drivers (presumably firmware) has been affected by the update.

TechToolPro reports no errors, and DiskWarrior is OK with it too. I am only seeing one swapfile in /private/var/vm, and nothing unexpected out of top. The fact that I get the plain blue screen (before the desktop appears properly) for 10 minutes or so suggests to me that it's not an application-based memory leak.

Starting up with apple-V shows me lots and lots of "System Starter: Waiting for local disks" messages. It looks and feels like the internal drive is limping along in 1st gear.

Anyone got any ideas? Thanks in advance.

cheers
RET

PS: EscapePod is the only hack running - no themes or window shades etc

[Stratus Fear]

There's definitely some kind of memory leak going on. I haven't thoroughly tested it, but the only things I've installed on my iBook since 10.2.6 was the new Safari update. Even before that, my RAM just doesn't hold up under .6. I NEVER used to get pageouts (I have 640MB) and now I get them all the time -- the free ram just slowly goes down over an hour or two. That's the sign of a memory leak right there. That's annoyed me enough where I have just been using Linux instead of X lately (that's not the only reason, I've been experimenting with Linux a lot as well).

[AirRon]

ret,

A simlar thing happened to me and it turned out that my ram had failed (and I was running OS X on 32MB on a Bondi blue iBook).
You probably already checked that, but it's just a thought.

Ronald

[ret]

Ron,

Thanks for that suggestion. TTP doesn't report any RAM problems, but hey, you never know. I'm currently running the repair permissions utility. It's identified a few things that are out of whack. I doubt that's the problem, but I might be lucky.

cheers
RET

[chris v]

Originally posted by ret:
Ron,

Thanks for that suggestion. TTP doesn't report any RAM problems, but hey, you never know. I'm currently running the repair permissions utility. It's identified a few things that are out of whack. I doubt that's the problem, but I might be lucky.

cheers
RET

Have you dumped the "Recent servers" cache? I've had my machines bog a bit on startup trying to find servers that weren't there. (I think Mac Janitor will clean that out for you.)

CV

[IamBob]

Here's my stuff to compare...

Code:
ps -auxcm

USER   PID %CPU %MEM      VSZ    RSS  TT  STAT STARTED      TIME COMMAND
ryan  1341   0.0 52.5   239544 137636  ??  S     9:00PM  17:00.10 System Events
ryan  1537   0.0  9.4   125864  24768  ??  S     7:07AM   1:09.72 Safari
ryan  1327   3.0  3.5    69748   9068  ??  Ss    8:59PM   2:36.02 Window Manager
ryan  1539   0.1  1.9    66444   5000  ??  S     7:47AM   0:01.35 Terminal
ryan  1339   0.0  0.9    92264   2288  ??  S     9:00PM   0:47.31 Finder
ryan  1328   0.0  0.7    73056   1808  ??  Ss    8:59PM   0:02.09 loginwindow
ryan   170   0.0  0.7    75492   1760  ??  Ss   Tue04PM   3:54.44 ATSServer
ryan  1337   0.0  0.6    69664   1672  ??  S     9:00PM   0:01.67 Dock
ryan  1338   0.0  0.6    64496   1612  ??  S     9:00PM   0:04.27 SystemUIServer
root   377   0.0  0.5    21880   1348  ??  S    Tue04PM   0:02.72 DirectoryService
root   312   0.0  0.4    28836   1072  ??  Ss   Tue04PM   0:03.83 coreservicesd
ryan  1540   0.0  0.3     5872    808 std  Ss    7:47AM   0:00.10 tcsh
root   106   0.0  0.3    16776    772  ??  Ss   Tue04PM   0:15.13 configd
ryan  1333   0.0  0.3    14724    736  ??  Ss    8:59PM   0:01.27 pbs
root   153   0.0  0.2    15256    584  ??  Ss   Tue04PM   0:01.68 SecurityServer
root   670   0.0  0.2    15220    504  ??  Ss   Tue09PM   0:04.22 lookupd
ryan  1379   0.0  0.2    14924    448  ??  S     9:37PM   0:00.09 AppleSpell
root   393   0.0  0.2    19136    420  ??  Ss   Tue04PM   0:09.27 AppleFileServer
[ ... ]

vm_stat

Mach Virtual Memory Statistics: (page size of 4096 bytes)
Pages free:                     1035.
Pages active:                  36525.
Pages inactive:                18040.
Pages wired down:               9932.
"Translation faults":       10181209.
Pages copy-on-write:           92223.
Pages zero filled:           4762867.
Pages reactivated:            163762.
Pageins:                       26181.
Pageouts:                      36117.
Object cache: 36121 hits of 76265 lookups (47% hit rate)

I don't have any haxies installed, I updated through Software Update (not a combo), 15" FP iMac/800/256/GF2MX...

What on earth does System Events do that it needs 52.5% of the memory? Killing it seems to fix the problem and doesn't adversely affect anything I use.

[Esquare]

Originally posted by IamBob:
I don't have any haxies installed, I updated through Software Update (not a combo), 15" FP iMac/800/256/GF2MX...

What on earth does System Events do that it needs 52.5% of the memory? Killing it seems to fix the problem and doesn't adversely affect anything I use.

Do you, by any chance, use Folder Actions? Those are handled by System Events...



EE

[eno]

Originally posted by IamBob:
Here's my stuff to compare...

Code:
vm_stat

Mach Virtual Memory Statistics: (page size of 4096 bytes)
Pages free:                     1035.
Pages active:                  36525.
Pages inactive:                18040.
Pages wired down:               9932.
"Translation faults":       10181209.
Pages copy-on-write:           92223.
Pages zero filled:           4762867.
Pages reactivated:            163762.
Pageins:                       26181.
Pageouts:                      36117.
Object cache: 36121 hits of 76265 lookups (47% hit rate)

Can anyone give any tips on how to interpret this info? For me, the man page is still a little cryptic... especially the entries for "Translation faults", and "Pages copy-on-write".

Code:
     Pages free
             the total number of free pages in the system.
     Pages active
             the total number of pages currently in use and pageable.
     Pages inactive
             the total number of pages on the inactive list.
     Pages wired down
             the total number of pages wired down.  That is, pages that cannot
             be paged out.
     Translation faults
             the number of times the "vm_fault" routine has been called.
     Pages copy-on-write
             the number of faults that caused a page to be copied (generally
             caused by copy-on-write faults).
     Pages zero filled
             the total number of pages that have been zero-filled on demand.
     Pages reactivated
             the total number of pages that have been moved from the inactive
             list to the active list (reactivated).
     Pageins
             the number of requests for pages from a pager (such as the inode
             pager).
     Pageouts
             the number of pages that have been paged out.

[malvolio]

Originally posted by Esquare:
Do you, by any chance, use Folder Actions? Those are handled by System Events...

Aha!
Thanks, Esquare! I have had 2 instances of the "runaway" System Events, did a simple "kill" each time and all was well, but I wondered just what it was I had killed.
Folder Actions need some work, it seems.

[suthercd]

Pages Free- basically the amount of free RAM that you have. No data stored there at the moment.

Pages Active- number of pages currently used by system and apps. These are paged out to disk last.

Pages Inactive- Inactive pages are segments of physical memory that have not been accessed within a given time frame. If these pages are modified, they must be paged out to disk. Inactive pages will be paged out to disk (virtual memory) as more memory demands are made, then once the page is unmapped of data, it is returned to the free page list. Inactive pages are paged out based on time since accessed- longest since referenced goes first.

If activity causes a page to be referenced, it go back to the top of the list to be paged out last. As physical memory is freed up, pageouts decrease, then cease at a given threshold. Pages are next reclaimed from the disk and put on the free list.

Pages Wired Down- Memory reserved for the operating system and is not susceptible to paging

Translation Faults- Not very useful info on an average system.

Pages Copy-on-Write- Say two apps share the same data. As one app changes that data, the old data is copied to a new location in RAM thus preserving it for the other app, and the new data is written separately. Each app uses the data that it "knows" about. The old data was copied on write.

Pages Zero Filled- Pages w/o data- all zeros. This number is cumulative since boot and continues to grow.

Pages Reactivated- Inactive pages that became active again. Cumulative since boot.

Pageins- pages read from disk to RAM. Pageins always happen at startup since the daata is on disk.

Pageouts- as you know by now, pages sent from RAM to disk.

Object Cache- Rates the efficiency of the cache process- if the data sought is in one of the caches that data can be quickly accessed. Higher the better.

As demands on memory increase, most physical memory is active, and the process of finding inactive pages is not possible. The system then may swap low priorty processes from RAM to disc or deactivate them entirely until memory is freed.

More info:
http://www.ixora.com.au/notes/virtual_memory.htm

HTH
Craig

[Sarc]

has anyone notices a huge memory leak while using bittorrent ?

[IamBob]

Folder Actions...ok, that makes sense.

I had putzed around with 'em a little forever ago and haven't used them since. I removed System Events from my login items (thought of looking about an hour ago ) and my performance woes be gone.

[eno]

Now with a little bit more uptime under the belt, I am up to more than 1.7 million page outs.

Code:
Pageouts:                    1739105.
Object cache: 495993 hits of 738800 lookups (67% hit rate)

Code:
7:44PM  up 5 days,  9:37, 4 users, load averages: 0.29, 0.47, 0.49

That still seems like an awful lot of pageouts to me, on a machine that has 1GB of RAM. I don't remember it getting that many a while back...

And right now I am only running about a dozen apps (Finder, Mail, Address Book, iCal, NetMonitor, Proteus, Terminal, JellyFiSSH, iTunes, Safari, Project Builder, Interface Builder, Aqua X-Chat, System Preferences, Preview) plus background processes.

And most of the day I've been out at work...

And in the minute it's taken me to type this message...

Code:
Pageouts:                    1739687.

I've added another 582 pageouts....

Seems like lots....

And unlike the others on this thread, it seems a general problem. I don't have System Events running, and I don't have any Folder Actions set. There doesn't seem to be any one memory hog in the list of apps....

The ouput of ps auxcm | head -5:

Code:
USER      PID %CPU %MEM      VSZ    RSS  TT  STAT STARTED      TIME COMMAND
enofile 18276  14.8 16.0  1440916 167728  ??  S    Thu05AM  46:38.62 Mail
enofile 24166  11.1 10.8   819184 113008  ??  S     7:11PM   4:02.62 Safari
enofile   179   5.4  5.7   805560  59564  ??  Ss   Sun10AM 312:29.71 Window Manager
enofile 24037  10.0  3.8   683164  40004  ??  S    12:12PM   5:29.03 iTunes

[Esquare]

Originally posted by eno:
Now with a little bit more uptime under the belt, I am up to more than 1.7 million page outs.

[...]

And right now I am only running about a dozen apps (Finder, Mail, Address Book, iCal, NetMonitor, Proteus, Terminal, JellyFiSSH, iTunes, Safari, Project Builder, Interface Builder, Aqua X-Chat, System Preferences, Preview) plus background processes.

I think a more systematic approach is in order. It is simple enough to 'kill' a few of these apps, to see if the paging-out stops. It might very well be a memory leak in one of these: start with the 3rd party ones.

EE

[suthercd]

To watch the memory management process, enter top -u in the Terminal. It is helpful to enter man top first to get information about the top command. In the man page, space bar scrolls down and exits or ctrl-c exits.

In the upper part of the window is a summary of data that is being discussed in this thread. Below is a list of processes running with the amounts of memory, current cpu use, threads by process and other info.

For those who have reported that their machine is experiencing excessive pageouts, etc.- how do you know that number is excessive?

Craig

[JoeyA]

FWIW, I've been experiencing some general weirdness with 10.2.6 as well. Since updating my hard drive has been grinding incessantly as if it's constantly paging out. On my system, a G4 Cube 450 with 1 GB of RAM, pageouts historically are few and far in between. Especially when using the same programs I've had open all the time in 10.2.5 without incident (Photoshop, Illustrator, GoLive, iTunes, Word and Safari).

In addition to the pageouts I also noticed two other things. Firstly, my cursor will jump erratically while my system slowly grinds to a halt. This gets worse and worse until I'm forced to close all my applications. During this time, CPU monitor reports heavy system processor usage while top reports the dock and Safari are eating about 20% of the processor a pop. Killing the dock correct the issue while closing an launching Safari will do the same.

The next thing I've noticed is that my dock will stop reporting which apps are open. The little black arrow indicating that a program has been launched simply vanishes. However clicking on the app's icon in the dock will switch you to that app happily humming along.

Perhaps this is all related somehow? I've really hadn't had many problems with OS X since the advent of Jaguar so this noticeable performance drop has been a real downer.

[wadesworld]

It's important to remember that pageouts are a normal part of the system.

No matter how much RAM you have, you're going to have pageouts.

The tricky part is determining when your pageouts become "too many".

Wade

[mrtew]

Originally posted by wadesworld:
It's important to remember that pageouts are a normal part of the system.

No matter how much RAM you have, you're going to have pageouts.

I won't tolerate them. I watch my memory monitor utility and when my Gig of RAM starts to get full I either restart or quit some apps. I only have to do that once every day or two and I never have any pageouts that way. I don't like them.

[eno]

Originally posted by Esquare:
I think a more systematic approach is in order. It is simple enough to 'kill' a few of these apps, to see if the paging-out stops. It might very well be a memory leak in one of these: start with the 3rd party ones.

EE

Code:
vmstat 1

Works pretty well for watching this in realtime. Much better than top -u.

One of my biggest memory hogs appears to be Mail.app. But I think that could be because I am running the httpMail bundle (for checking hotmail) and that often causes thrashing. As soon as I disconnect from hotmail the thrashing stops. And I see zombie Mail threads owned by root even after I quit Mail.app.

[eno]

Originally posted by mrtew:
I won't tolerate them. I watch my memory monitor utility and when my Gig of RAM starts to get full I either restart or quit some apps. I only have to do that once every day or two and I never have any pageouts that way. I don't like them.

I don't like excessive pageouts, but what you've just said has to be the most idiotic piece of misinformed dogmatism I've ever heard on this forum (and I've heard lots!)

But by the same token, you're welcome to be stupid if you really want to be.

By all means, go on rebooting every day in order to stop pageouts from happening; and I'll enjoy the pleasure of leaving my machine on 24/7 with all the apps I need constantly running at available to me at a moment's notice.

[CarpetFluff]

Not to sure about all the technobabble going on here but since upgrading to 10.2.6 I've also had my machine turning into an unresponsive hard drive grinding slug. It seems to occur very soon after running any big application such as Photoshop.

I first notice that the system has slowed down because the dock is a little sticky, even if I quit everything.

I have a DP533 with 640Mb of RAM 60Gb of disk space remaining, no 'haxies' installed, I don't usually run anything out of the ordinary, Mail, iTunes, Adium, Safari.

[King Bob On The Cob]

9 Swapfiles?!?

[Jim Stoneburner]

OK, the mysterious delay in launching apps gradually became worse, making system intolerable after 48 hours uptime.

Symptoms:

When an app is launched, there is a long pause with a beachball while the disk accesses (usually after menubar and window are drawn). Access is not rapid-fire thrashing (like VM read-write), but more like the sound of a search from the Finder (as when accessing a submenu for the first time)... not as loud or rapid a noise.

I see the same long access time when accessing a submenu full of files (not just folders) in Fruitmenu for the first time. Normally after a reboot, the submenus load in a second or two.

(If I let more time pass since the last reboot, the delays get much worse.... For example, a launched app's name may not even appear in the menubar for 30 seconds, followed by yet more delays as it draws the window and the beachball continues to spin. If another app is brought forward, the launched app cannot be brought forward until the beackball stops.)

Fixes attempted:

The following items were tried with no change:

• Quit all applications.
• Quit PTHPasteboard and PTHClock.
• Killed iChat Agent, escapepod.
• Repaired permissions,
• Updated prebindings, and
• Deleted user and system caches, including VM swap files, using Cocktail.
• Quit comapcpowerchute.
• Quit and relaunched Finder.
• Killed USB Monitor.

... No change.

(Also, as stated earlier, DiskWarrior, fsck, and several of the above had already been tried prior to rebooting.)

Diagnostic data:

Code:
Processes:  48 total, 3 running, 45 sleeping... 129 threads            12:10:46
Load Avg:  1.09, 0.76, 0.57     CPU usage:  77.5% user, 22.5% sys, 0.0% idle 
SharedLibs: num =  126, resident = 39.5M code, 2.74M data, 12.2M LinkEdit
MemRegions: num = 4415, resident = 70.7M + 5.98M private, 70.9M shared
PhysMem:  73.5M wired,  161M active,  294M inactive,  529M used,  111M free
VM: 1.48G + 73.1M   27315(0) pageins, 6367(0) pageouts       

  PID COMMAND      %CPU   TIME   #TH #PRTS #MREGS RPRVT  RSHRD  RSIZE  VSIZE
 6175 top          9.2%  0:01.31   1    14    17   252K   328K   548K+ 13.6M
 5965 Conduit Ma   0.0%  0:28.94   3    78   147  2.19M  8.37M  11.1M  49.6M
 5743 tcsh         0.0%  0:00.12   1    10    15   344K   592K   792K  5.73M
 5729 Console      0.5%  0:31.01   2    66   183  2.49M  8.30M  17.2M  50.2M
 5684 Terminal    75.4%  0:39.92   4    72   206  2.54M  8.54M  17.4M  50.9M
 5448 perl         0.0%  0:01.34   1     9    14   144K   304K   388K  1.27M
 5447 perl         0.0%  0:00.90   1     9    14   168K   304K   408K  1.27M
 5446 perl         0.0%  0:03.35   1     9    14   164K   304K   404K  1.27M
 5445 perl         0.0%  0:01.43   1     9    14   144K   304K   384K  1.27M
 5444 perl         0.0%  0:00.61   1     9    14   144K   304K   384K  1.27M
 5438 Overseer     6.3%  1:40.22  14    91   237  5.12M+ 9.61M  20.8M+ 55.6M+
 5294 lookupd      0.0%  0:02.32   2    33    45   344K   472K   720K  14.9M
 1359 AppleSpell   0.0%  0:00.12   1    16    21   472K  1.17M   692K  14.6M
  682 Tex-Edit P   0.5% 12:28.95   2    86   412  16.4M  22.2M  25.4M  85.3M
  534 SecurityAg   0.0%  0:01.72   3    82   152  1.66M  4.43M  2.76M  47.7M
  409 USB Monito   0.0%  0:17.81   2    87   139  1.51M  4.48M  3.48M  46.4M
  405 Transport    0.5%  1:21.08   3    87   139  1.57M  4.97M  3.63M  46.1M
  401 Finder       0.5%  3:00.78   2    99   285  11.2M  29.3M  22.2M  88.3M
  400 SystemUISe   0.0%  0:40.24   3  4972   208  5.18M  4.70M  6.49M  50.8M
  399 aped         0.0%  0:00.86   1    27    22   144K   696K   384K  14.0M
  398 Dock         0.0%  5:31.67   3   133   232  1.42M  8.65M  3.42M  52.0M
  392 pbs          0.0%  0:02.01   2    27    28   588K   900K   968K  14.4M
  387 DirectoryS   0.0%  0:03.03   3    54   143   620K  3.84M  1.44M  21.4M
  376 loginwindo   0.0% 68:15.49   5   178   225  3.05M  7.16M  6.49M  41.7M
  365 automount    0.0%  0:00.09   2    22    22   124K   388K   328K  14.7M
  357 coreservic   0.0%  0:06.57   3    92   125  2.80M  16.7M  18.8M  39.0M
  346 nfsiod       0.0%  0:00.00   1     8    13     0K   300K    28K  1.27M
  345 nfsiod       0.0%  0:00.00   1     8    13     0K   300K    28K  1.27M
  344 nfsiod       0.0%  0:00.00   1     8    13     0K   300K    28K  1.27M
  343 nfsiod       0.0%  0:00.00   1     8    13     0K   300K    24K  1.27M
  331 autodiskmo   0.0%  0:00.65   1    38    21   160K   692K   252K  14.6M
  306 cron         0.0%  0:00.91   1     8    18    52K   332K   104K  13.8M
  304 cupsd        0.0%  0:36.64   1     8    49  4.60M   912K  4.75M  19.6M
  297 inetd        0.0%  0:00.01   1     8    14    20K   312K    32K  1.28M
  277 netinfod     0.0%  0:08.14   1     7    33   324K   432K   464K  1.91M
  260 crashrepor   0.0%  0:00.00   1    14    15    32K   300K    48K  13.3M
  256 mDNSRespon   0.0%  0:00.56   2    29    26   652K   352K   836K  14.2M
  247 syslogd      0.0%  0:01.41   1     7    14    68K   312K   116K  1.28M
  176 Window Man   4.0%  2:11:01   3   268   457  7.79M  22.1M  28.4M  66.4M

[Jim-Stoneburner:~] jds% vm_stat
Mach Virtual Memory Statistics: (page size of 4096 bytes)
Pages free:                    28410.
Pages active:                  41287.
Pages inactive:                75317.
Pages wired down:              18826.
"Translation faults":       69667586.
Pages copy-on-write:          161077.
Pages zero filled:          61672501.
Pages reactivated:            275914.
Pageins:                       27316.
Pageouts:                       6367.
Object cache: 101413 hits of 517442 lookups (19% hit rate)

More data in next message...

[Jim Stoneburner]

In Terminal, running "ls -l /private/var/vm" returned no result.

Code:
[Jim-Stoneburner:~] jds% ps -auxcm
USER   PID %CPU %MEM      VSZ    RSS  TT  STAT STARTED      TIME COMMAND
jds    176   1.4  4.1   590948  27184  ??  Ss   Thu10AM 131:34.68 Window Manager
jds    682   3.0  4.1   612780  26956  ??  S    Thu08PM  12:52.21 Tex-Edit Plus
root   357   0.0  3.0    39864  19352  ??  Ss   Thu10AM   0:07.09 coreservicesd
jds   5438   2.5  2.3   580292  15196  ??  S    11:58AM   2:51.56 Overseer
jds   5684   4.1  1.8   576576  11680  ??  S    12:02PM   0:43.91 Terminal
jds   6831   0.0  1.7   575140  11404  ??  S    12:21PM   0:28.72 LaunchCFMApp
jds   6507   0.0  1.7   592864  10964  ??  S    12:16PM   0:30.10 Finder
jds   6756   0.0  1.4   575804   9288  ??  S    12:20PM   0:29.85 System Preferenc
jds   5729   0.0  1.4   575720   8928  ??  S    12:03PM   0:31.55 Console
jds    400   0.0  1.0    52028   6664  ??  S    Thu10AM   0:40.37 SystemUIServer
jds    376   0.0  1.0    42344   6644  ??  Ss   Thu10AM  68:15.89 loginwindow
root   304   0.0  0.7    20044   4868  ??  Ss   Thu10AM   0:36.75 cupsd
jds    171   1.5  0.6    42260   4088  ??  Ss   Thu10AM  11:48.88 ATSServer
jds    405   0.0  0.6   571476   3700  ??  S    Thu10AM   1:21.28 LaunchCFMApp
jds    398   0.0  0.5    53284   3528  ??  S    Thu10AM   5:36.20 Dock
jds    534   0.0  0.4    48860   2824  ??  S    Thu12PM   0:01.73 SecurityAgent
root   387   0.0  0.2    21868   1476  ??  S    Thu10AM   0:03.05 DirectoryService
jds    392   0.0  0.1    14724    936  ??  Ss   Thu10AM   0:02.01 pbs
root   109   3.0  0.1    16708    840  ??  Ss   Thu10AM   5:40.03 configd
root   256   0.0  0.1    14516    832  ??  Ss   Thu10AM   0:00.56 mDNSResponder
jds   5743   3.0  0.1     5872    800 std  Ss   12:03PM   0:00.16 tcsh
root  5294   0.0  0.1    15220    724  ??  Ss   11:34AM   0:03.53 lookupd
jds   1359   0.0  0.1    14924    692  ??  S    10:04AM   0:00.12 AppleSpell
root   139   0.0  0.1    15240    492  ??  Ss   Thu10AM   0:00.52 SecurityServer
root   277   0.0  0.1     1960    464  ??  Ss   Thu10AM   0:08.72 netinfod
jds   5447   0.0  0.1     1300    408  ??  Ss   11:59AM   0:01.83 perl
jds   5446   0.0  0.1     1300    404  ??  Ss   11:59AM   0:07.01 perl
jds   5448   0.0  0.1     1300    392  ??  Ss   11:59AM   0:02.85 perl
jds    399   0.0  0.1    14312    392  ??  S    Thu10AM   0:00.89 aped
jds   5445   0.0  0.1     1300    384  ??  Ss   11:59AM   0:02.77 perl
jds   5444   0.0  0.1     1300    384  ??  Ss   11:59AM   0:01.26 perl
root    51   0.0  0.1    15656    368  ??  Ss   Thu10AM   0:03.40 kextd
root   365   0.0  0.1    15008    332  ??  Ss   Thu10AM   0:00.09 automount
root  6948   0.0  0.0     1360    308 std  R+   12:23PM   0:00.00 ps
root   331   0.0  0.0    14968    264  ??  Ss   Thu10AM   0:00.66 autodiskmount
root     1   0.0  0.0     1308    176  ??  Ss   Thu10AM   0:00.06 init
root     2   0.0  0.0     1844    164  ??  Ss   Thu10AM   0:02.72 mach_init
root   247   0.0  0.0     1308    120  ??  Ss   Thu10AM   0:01.41 syslogd
root   306   0.0  0.0    14124    108  ??  Ss   Thu10AM   0:00.93 cron
root    78   0.0  0.0     1316     64  ??  Ss   Thu10AM   0:00.01 dynamic_pager
root    73   0.0  0.0     1292     60  ??  Ss   Thu10AM   0:59.75 update
root   260   0.0  0.0    13584     48  ??  Ss   Thu10AM   0:00.00 crashreporterd
root   297   0.0  0.0     1308     32  ??  Ss   Thu10AM   0:00.01 inetd
root   344   0.0  0.0     1296     28  ??  S    Thu10AM   0:00.00 nfsiod
root   345   0.0  0.0     1296     28  ??  S    Thu10AM   0:00.00 nfsiod
root   346   0.0  0.0     1296     28  ??  S    Thu10AM   0:00.00 nfsiod
root   343   0.0  0.0     1296     24  ??  S    Thu10AM   0:00.00 nfsiod
root  6947   0.0  0.0        0      0 con- Z    31Dec69   0:00.00 ps
root  6946   0.0  0.0        0      0 con- Z    31Dec69   0:00.00 vm_stat
root  6945   0.0  0.0        0      0 con- Z    31Dec69   0:00.00 tail
root  6944   0.0  0.0        0      0 con- Z    31Dec69   0:00.00 hostinfo
root  6943   0.0  0.0        0      0 con- Z    31Dec69   0:00.00 df
[Jim-Stoneburner:~] jds%

Killing Window Manager caused an OS X blue screen and reloading of Dock, Finder, etc. All remaining apps had been quit by this. AFter this, the delay was gone.

Next steps: Obtain advice on above, then try rebooting and 2 days running with all third-party stuff disabled. Perhaps 10.2.6 disagrees with one of them?

Question remains: why does this delay occur and progressively get worse with time?

[suthercd]

Jim-

Thanks for the update. Read through the list of changes you have made- did you ever create a new user and login to that account?

Craig

[Jim Stoneburner]

Originally posted by suthercd:
did you ever create a new user and login to that account?

Craig

Not yet. Once the symptoms returned, I tried by eliminating apps and process without rebooting. Failing to resolve it, I took data to look for signs of what's up.

Next, I'll try trimming the system of third party stuff.

If those "exclusionary" tests don't reveal anything, then a new user is next to try.

If a new user continues to have the problem, it would seem that something in the system is damaged, something my various repairs can't get to. Either that, or it is a bug somewhere that my system setup manages to trigger (particular types of USB devices, pref settings, etc.).

(I get annoyed when people say "my 10.2.6 systems are fine" as if that means anything. Most bugs affect a minority of users, for specific reasons. That kind of statement simply implies that there isn't a bug which affects 100% of systems... that I knew already!)

[malvolio]

Originally posted by Jim Stoneburner:
In Terminal, running "ls -l /private/var/vm" returned no result.

You mean nothing was listed? That is highly unusual, and almost certainly indicates a significant problem. Have you ever tried to set up a dedicated swapfile partition?
Can you beg, borrow, or steal (no, don't steal) a DiskWarrior CD to repair your directory? Failing that, my best advice is to back up your data and reinstall OS X.

[Jim Stoneburner]

You mean nothing was listed?

You must've missed what I wrote. I had just deleted system caches, including VM swap files, using Cocktail. No surprise to find none listed.

[malvolio]

Even with no swapfiles,you should see the following:
[Sat 9:57:26pm vm]% ls -al
total 0
drwxr-xr-x 3 root wheel 102 May 11 01:16 .
drwxr-xr-x 20 root wheel 680 May 24 12:02 ..
drwx--x--x 10 root wheel 340 May 20 16:02 app_profile

Were those items present?

[Jim Stoneburner]

Originally posted by malvolio:
Even with no swapfiles,you should see the following:
[Sat 9:57:26pm vm]% ls -al
total 0
drwxr-xr-x 3 root wheel 102 May 11 01:16 .
drwxr-xr-x 20 root wheel 680 May 24 12:02 ..
drwx--x--x 10 root wheel 340 May 20 16:02 app_profile

Were those items present?

That is a different command. Seriously, I think you are taking me on a dead-end tangent.

Here are three different commands in a row, below. Keep in mind I had just deleted all system caches and killed Window Manager. All processes were killed and some(?) relaunched. I am not seeing any slowdown right now.

Code:
[Jim-Stoneburner:~] jds% ls -l /private/var/vm
[Jim-Stoneburner:~] jds% vm_stat
Mach Virtual Memory Statistics: (page size of 4096 bytes)
Pages free:                     2359.
Pages active:                  57367.
Pages inactive:                85235.
Pages wired down:              18879.
"Translation faults":       86038512.
Pages copy-on-write:          197037.
Pages zero filled:          76776575.
Pages reactivated:            325130.
Pageins:                       34313.
Pageouts:                       6399.
Object cache: 121483 hits of 544574 lookups (22% hit rate)
[Jim-Stoneburner:~] jds% ls -al
total 2712
drwxrwxrwx   28 jds   staff      952 May 24 12:15 .
drwxrwxr-t    6 root  wheel      204 Mar  2 22:22 ..
-rw-rw-rw-    1 jds   staff        3 Dec 11 00:50 .CFUserTextEncoding
-rwxrwxrwx    1 jds   staff    15364 May 24 12:15 .DS_Store
-rw-rw-rw-    1 jds   staff  1351680 Jun 26  2002 .FBCIndex
drwxrwxrwx    3 jds   staff      102 Jun 26  2002 .FBCLockFolder
-rw-r--r--    1 jds   staff        0 Feb 14 11:53 .MCXLC
drwxrwxrwx    2 jds   staff       68 Nov 16  2002 .MacOSX

... and so on.

I think you may also have missed the fact that I ran DiskWarrior 3.0 before and after the update, and do so regularly.

[ret]

JoeyA and Jim,

I'm glad to see I'm not alone. I'm having the same problem with the constantly grinding disk drive, but it manifests itself well before the desktop actually appears. Rebooting the machine can take 20 minutes or more - as mentioned earlier when booting with apple-v down I see a lot of "waiting for local disks" and "waiting for application services" messages.

I have run DW, TechTool and defragged the HD (when I ran out of other ideas), none of which has made the slightest difference.

What is concerning me now is that in booting OS9 to run those tools it still took just as long to start-up, which suggests to me that it's either:
hardware problem, or
low-level disk driver problem
I'm not filled with joy by either of those alternatives....

RET

TiPB 500, 512MB Ram, 20GB Int HD