heh. regrets. here, all the entries are logged by the...

windowserver !!

but, the [pid] notation applies to most logs.

229 is my WindowServer pid (is that arbitrary or universal), yet 5941 pid doesn't appear for any of the apps I've mentioned (does that also dynamically change day to day).

As far as I know the pid get assigned once the process gets started, therefore by starting the application or when the computer boots. as long as the process runs it will have the same pid.

Wild Windowserver fs_usage with Clean install!
 

Using Panther on my eMac at work, the command
sudo fs_usage continues to show 60-80 lines per second of

09:34:18 CACHE_HIT 0.000028 WindowServer
09:34:18 CACHE_HIT 0.000008 WindowServer
09:34:18 CACHE_HIT 0.000007 WindowServer
09:34:18 CACHE_HIT 0.000007 WindowServer

and I have yet to trace down the reason why. Nothing interesting (any different than those errors listed in this thread) shows up in the windowserver.log, and not nearly so frequent at that.

My computer at home does not show this extreme windowserver activity under sudo fs_usage, but just shows one item per second (usually cupsd). The installations are the same, with the only difference being the network connection and that the eMac can use Quartz Extreme while my iMac at home cannot. Starting under a Safe Boot silences these messages.

Finally, I took the extreme step of erasing and re-installing Panther. Before I installed anything at all, I checked the results of sudo fs_usage and it was still the same...60-80 lines per second of the above. Similarly after all Apple Updates (and no personal installations) had been applied.

Can anyone else verify this on
1) An early vintage eMac
or
2) A Macintosh with Quartz Extreme?

PS: I'm not asking you to verify what happens after a clean install, just what the results of 'sudo fs_usage' are. :-)

Requests:

1) Could someone with an early eMac and Panther please run "sudo fs_usage" in Terminal and let me know if you see a stream of lines like:

09:34:18 CACHE_HIT 0.000028 WindowServer
09:34:18 CACHE_HIT 0.000008 WindowServer
09:34:18 CACHE_HIT 0.000007 WindowServer
09:34:18 CACHE_HIT 0.000007 WindowServer


2) Could someone whose machine has Quartz Extreme please run "sudo fs_usage" in Terminal and let me know if you see a stream of lines like those above.

Control-C will exit the command.

Thank you.

I've just read in another thread that Quartz Extreme is disabled in Safe Mode. The streams of 60-80 lines per second of lines like these

09:34:18 CACHE_HIT 0.000028 WindowServer
09:34:18 CACHE_HIT 0.000008 WindowServer
09:34:18 CACHE_HIT 0.000007 WindowServer
09:34:18 CACHE_HIT 0.000007 WindowServer

when viewing fs_usage *do not* occur when my eMac is started in Safe Mode and they *do not* occur on my summer 2001 iMac (which does not permit QE). Perhaps this is a coincidence, but I cannot check it alone.

Could someone with Quartz Extreme please run

sudo fs_usage

in Terminal while l little else is going on and tell me whether or not you get streams and streams of the lines like those 4 above (around 70 per second).

I'd like to know if it is a matter with my machine or if it is a QE thing.

Please.

I now have a report from another original model eMac 700 MHz of the same activity (with a fresh installation of Panther on it). That is, about 70 lines per second of lines like these

09:34:18 CACHE_HIT 0.000028 WindowServer
09:34:18 CACHE_HIT 0.000008 WindowServer
09:34:18 CACHE_HIT 0.000007 WindowServer
09:34:18 CACHE_HIT 0.000007 WindowServer

when runnning 'sudo fs_usage'. So, it does not seem to be something only with my eMac. If I start up from OS X 10.2.8 (still on another partition), it this does not happen.

I am guessing that it is a QE and Panther problem, but I can't verify this until I can get some results with a non-eMac with Panther and QE.

Hi guys.

Located this thread, however it never seemed to reach a conclusion?

Can you help how to solve the repeated logs noted in the command:
sudo fs_usage

I have the same problems with hundreds of lines between written every few seconds. :(

I'm assuming you are talking about the CACHE_HIT lines from WindowServer
Why is this a problem? These lines are not being written to any file, hence disk space usage is not an issue. They merely reflect the fact that the WindowServer process is reading some data from a file (and that this file is cached in memory so the read is actually from RAM).

Yes. I was referring the logs, an example as below:
11:04:02 CACHE_HIT 0.000011 WindowServer
11:04:02 CACHE_HIT 0.000010 WindowServer
11:04:02 CACHE_HIT 0.000009 WindowServer
11:04:02 CACHE_HIT 0.000005 WindowServer
11:04:02 CACHE_HIT 0.000004 WindowServer
11:04:02 CACHE_HIT 0.000004 WindowServer

Hmm. Well interesting question? Maybe you are right and it is not a problem? Although another user in this topic noted that he only had a handful of logs and not thousands every few seconds. Could you please run sudo fs_usage and let me know the output please.

It does not sound like just RAM memory activity. I can hear the harddrive working.

Everything works ok, and enough time has gone by that I have come to accept it, but my original concern was that I was losing cpu cycles to something that did not need to be happening. To a greenhorn like me, 70-80 hits each second seems like a lot, and like it might get in the way of processes you want to be happening. Also, if it shouldn't be happening, then something could be wrong that will sneak up and bite you later. Perhaps SML has similar concerns. That having been said, if you watch fs_usage, the WindowServer hits seem to move out of the way when you have other processes active.

Exactly right macmath. I just want to make sure my new powerbook is running at its best performance and not wasting cpu and RAM and harddrive speed with an error proble.

Perhaps if some other users can note the output of the command, we may find it is perfectly normal?!

SML, if you can afford to increase your RAM from 256 to 640 or 768MB, I think that your PB would run circles around what it does now (and no doubt it does well now).

I didn't see any entries from WindowServer when I ran fs_usage for about a minute just now. But I don't have Quartz Extreme - I'm on an iBook from fall 2001.

If you hear drive activity, that is not from the CACHE_HIT's. It is something else. First thing to check is the pageins/pageouts listed under the System Memory tab in Activity Monitor. Lots of pageouts indicates you need more RAM.

There were heaps of constant page in/outs on the pmTool.

The system memory is pretty full with two or three apps running. Page ins/outs are 28047/19620 with the system turned on for a few hours. Seems constant though and not increasing - although I can still hear the hard drive activity.

So it sounds like something other than virtual memory is hitting the drive. If you are seeing a lot of CACHE_HIT's when you do fs_usage, it might be useful to filter them out so you can better see what is happening. You could do this with:

sudo fs_usage | grep -v CACHE_HIT

If there is something else you want to filter out, just add another grep -v for that at the end. E.g. to filter out all lines with "foo bar" in them:

sudo fs_usage | grep -v CACHE_HIT | grep -v "foo bar"

It is easy to see what is happening just from the fs_usage output. There is nothing else apart from the following .... and with the odd usual app process such as cups every minute or two. Here is the output with further detail - command
"sudo fs_usage -w"

16:48:05.077 CACHE_HIT A=0x118f6000 0.000004 WindowServer
16:48:05.077 CACHE_HIT A=0x118f6000 0.000004 WindowServer
16:48:05.077 CACHE_HIT A=0x118f6000 0.000004 WindowServer
16:48:05.077 CACHE_HIT A=0x118f6000 0.000004 WindowServer
16:48:05.078 CACHE_HIT A=0x118f6000 0.000004 WindowServer
16:48:05.078 CACHE_HIT A=0x118f6000 0.000004 WindowServer
16:48:05.078 CACHE_HIT A=0x118f6000 0.000004 WindowServer

Agree that more RAM will be great for my PB, but that is not the problem in this situation. Another user running 1.2Gb of RAM has the same problem.

Sorry, I've only just come across this thread. I had found the exact same problem on a friends ibook. After removal of the HPDeskjet driver the quite loud disk activity stopped.

Running SnapzPro?
 

I had this problem -- repeated errors logged to the windowserver.log.

Turns out the problem is a known (to Ambrosia) problem caused by an interaction between SnapzPro and any other application (for me, Codetek Desktop) that creates a weird (to WindowServer, invalid) window.

Quitting SnapzPro solved it for me.

Details here.