RichB

I'm running leopard here (unfortunatley ) and I've got two Lacie USB drives. One 500GB used for Time Machine and another 320GB for 'stuff'
I now can't use the 320 as with it plugged in, it regularly disappears from the Desktop, causes the machine to grind to a halt. If I unmount the 320 the machine fires back into life again.

Something cr*p is going on with Leopard, running very little last night, just downloading some pics of my camera and the machine was running so slow, (3-5 minutes to open sys prefs)

I am using a Dual 2.5 G5 - OK, not Intel but it's dual processor and 64 bit.
It currently has 2gb in it which has always been fine, the only process showing in activity monitor as 64bit is httpd strangely.

Order another 2gb of Ram to see if it helps but at the moment I am really unimpressed with this pile of poo.

Marcus - Do you have any idea why in activity monitor httpd processes are created each time you open a new browser window and access a page off your built in webserver.
Also, do you think it is strange that each instance of an httpd process reports 6.59 GB of virtual memory - Seems odd to me.
The other day when the machine was running very slow, killing those processes sorted it out.

Rich

ScoobyDoo555

I'd heard some rumours of issues like yours Rich (not much help, I know! )
I'm currently sticking with Tiger for the time being - it seems to me (and it's just IMHO) that Apple rushed this out, and took their collective eye off the ball.

But it does seem that Apple have got very resource hungry in their latest releases.... just installed Logic Studio 8. This install was just under 39Gig!!

In the meantime, will watch with anticipation - and might cross Leopard off my Xmas list!! Hope you get it sorted
Dan

Markus

It's using HOW MUCH vm? That's a fair bit (understatement of the century I know).

Ok, when you fire up Activity Monitor, set the process list popup to "All Processes" in the search field, type in "htt" (without quotes), what do you get back (screenshot might be nice). I ask as at present with seven browser windows open I only have two httpd processes listed.

The key thing here is also the "user" column, what is that reporting for the username? my two entries are root and "_www".

I've been using Safari to open many other windows whilst reading this and I do not see any other additional processes.

Of further note I do have web sharing enabled, and when I disable it those two processes are quit. Which makes sense as I was going to say that httpd is the http daemon and as such it should only be there if you have web sharing enabled.

Thoughts? Well, do you have web sharing enabled? Do you need it enabled is probably a better question, if not, disable it. If you do, then temporarily turn it off and see if you get the same results, if so then something else could be launching the httpd daemon.

To find out what launched something, select the item in Activity Monitor and then click "Inspect". You'll see at the top of the window that opens a "Parent Process" item, which should display the name of the process that launched the item.

My second httpd process, with a user of _www has a parent process of httpd (2373) - the number being the process id. It's a high number as I stopped and started web sharing to see if the items quit, which they did, anyway I digress. So, click on the blue link bit and it will take you to the parent process, in my case the other httpd process (user of root). That item will also have a parent process and here I have launchd, which I think is the launchdaemon handler, clicking that items parent process takes me to kernel task, and clicking that goes to kerenel task 0 and that is as high as I can go, signifying that's the real parent process.
That layout given is quite normal and what you'd expect.

I'm guessing you may see something different. If you have PHP, PERL or other things that apache might use that it could spawn other processes, but not when you open additional browser windows, Safari should not need to do that at all.

I am of course presuming here that you're using Safari and not another browser. I've just tried Firefox 2.0.0.9 and I still don't see the problem you do.

Simon C

I've noticed a few funny things running 10.5 as well.


Things like firefox and dreamweaver not shutting down. Taking its merry little time switching between pages in dreamweaver and getting very bogged down doing the most simple of photoshop tasks like cropping.

As time machine won't do what I want, I am seriously thinking about regressing back to Tiger and being done with it.

JackClark

The only real benefit I've found with Leopard is the Mosaic screensaver, I won't go back to Tiger for that reason alone.

RichB

Marcus, I build websites on my Mac so need the web sharing.
Users are root and _www
Yes I'm using Safari in this instance.

I've used Dreamweaver before but since saving any file I had open as a blank file when it crashes I've gone right off it

When I open a new window, I am hitting a site on my machine,served by PHP and (most likely) serving data from a mysql DB so that might be the reason for the multiple httpd processes.

The biggest problem is the external disk. Just sat here for about 10 minutes waiting to see if the machine would come back to life which it didn't, turned off my 320 GB external and suddenly, things printed, windows close, all the mouse clicks I'd done were actioned.
Really annoying.

Now I've just started all my sharing.


Now I've gone to a site on my machine, same page 5 times in new tabs and here's the result.


I don't believe it's using 6.5GB of VM...
Any thoughts?

I tried turning off some modules in the php httpd.conf but reluctant to turn too much off, not sure what I really need... Wouldn;t think the default settings would be that bad... The only thing I switched on was php5!

Markus

Ok, if you're opening multiple windows and the default URL they access is your local site that would cause a few more httpd processes to spawn, and that is normal. When you close the windows they will keep the connections alive so they can be reused. The number you're seeing is similar to what i was seeing when testing things out.

There should be something in a config file, possibly httpd.conf (think it's in /etc/ ) which has something about the number of connections and the keepalive time which keeps them active when not being used, so you could change that. The VM being used does seem a little high, but it seems the same for all of the connections so I'd say that VM wise it's a shared size rather than 7GB for each connection, otherwise you'd be looking at 35GB of space being taken up. It could also be just an allocated value, as in, it could take UP to 7GB, but at the moment it's only taking say 100MB.

I don't think this problem is related to the drive issue, which is certainly an annoying nay frustrating issue. So what could be causing this. Ok, here's what I want you to do, fire up Activity monitor, set it to view all processes and click on the CPU column heading so it'll show the highest CPU process at the top of the list. Turn on the drive and wait for the dreaded slowdown. Then have a look at activity monitor and see what the first couple of processes are. This should, hopefully, give us a clue as to what the hell is knocking the machine for six.

Other things to try would be to switch the USB port the drive is attached to and see if that makes a difference, but you've probably already tried that.

Another thing to try is to reboot into safe mode (reboot, as soon as you hear the chime hold down and keep held down the shift key. When the login dialog appears it should say Safe Boot on it) Enter name and password, hold down the shift key and click the Login button, keep shift held down until you're at the desktop. That will stop all login items from running.

Turn on the drive, see if you can replicate the problem. Now some processes won't be running due to safe boot, ie; most startup items, and things that would auto launch at login won't be running either. This is good in a way in that if we don't see the problem we can start to look at what runs at startup and login.

The simple place to start is with login items. Go to: System Preferences -> Accounts -> Your Account -> Login Items
Have a look at what is listed, take a screenshot of the list. Now you have choice of what to do. you can remove all of the items, reboot and login and see if the problem still exists, or you can remove the items one at a time, rebooting between each removal to see if we can narrow down which one of the buggers is causing a problem.

Do you have any disk monitoring software, for example, TechTool Pro has a utility that will check the status of disks, turn that off and see if you get the problem.

f1_fan

iTrader: (9)

Rich,

I'm not a Mac expert but let's face it OS X is basically a *nix variant behind the scenes (however much the Mac fans like to hold it up as something magical) and if I had a Linux/Unix webserver that was grabbing virtual memory of 7GB for each httpd process I would be worried. Either it is being reported incorrectly or it doesn't mean what it says it does or something is very wrong.

The spawning of new processes is normal and if they had a sensible footprint you wouldn't notice them. Have you tried turning off PHP5 and then see if that cures the issue. If so, or anyway atually,is it possible to reinstall or rebuild the webserver under the new OS upgrade?

P1_BEN

Not sure but you could try disabliing these directive's if enabled, again in the httpd.conf file if you have found that. The fact it mentions having other file systems mounted which if you have external disks plugged in you do have.. Also comments on issues when multiple CPU's are involved.

Memory-mapping
In situations where Apache 2.x needs to look at the contents of a file being delivered--for example, when doing server-side-include processing--it normally memory-maps the file if the OS supports some form of mmap(2).

On some platforms, this memory-mapping improves performance. However, there are cases where memory-mapping can hurt the performance or even the stability of the httpd:

On some operating systems, mmap does not scale as well as read(2) when the number of CPUs increases. On multiprocessor Solaris servers, for example, Apache 2.x sometimes delivers server-parsed files faster when mmap is disabled.

If you memory-map a file located on an NFS-mounted filesystem and a process on another NFS client machine deletes or truncates the file, your process may get a bus error the next time it tries to access the mapped file content.

For installations where either of these factors applies, you should use EnableMMAP off to disable the memory-mapping of delivered files. (Note: This directive can be overridden on a per-directory basis.)

Sendfile
In situations where Apache 2.x can ignore the contents of the file to be delivered -- for example, when serving static file content -- it normally uses the kernel sendfile support the file if the OS supports the sendfile(2) operation.

On most platforms, using sendfile improves performance by eliminating separate read and send mechanics. However, there are cases where using sendfile can harm the stability of the httpd:

Some platforms may have broken sendfile support that the build system did not detect, especially if the binaries were built on another box and moved to such a machine with broken sendfile support.

With an NFS-mounted files, the kernel may be unable to reliably serve the network file through it's own cache.

For installations where either of these factors applies, you should use EnableSendfile off to disable sendfile delivery of file contents. (Note: This directive can be overridden on a per-directory basis.)

RichB

Regarding the disk issues, I've found a thread on Apples discussions where a load of people are having the same problem.
Apple - Support - Discussions - leopard hangs - have to power off - ...

RE the VM - I tried turning off PHP in httpd.conf and still got the same multiple processes, perhaps this is normal and irrelevant here.
The real problem is why the disk causes the hang.

Markus

Ok, so I'm guessing the last post on there (rjb something or other) is you.
I'm guessing ACD = Apple Cinema Display, if not, what is it. Have I read it correctly that you have a usb cable going from the front USB port on the Tower to the ports on the rear of the display and then you have the 7-port USB hub connected to the ports on the display and the keyboard is attached to one of the ports on the hub.

If so, here's what I'd try: Remove the connection between the display and the computer. Plug the keyboard directly into the front of the machine. Basically we're removing the display and the hub from the USB chain. See if you still get problems.
If you don't, then that is something. I'd then plug a USB cable from the computer (front) to one of the USB ports on the monitor, then stick the keyboard into the other port on the monitor. See if problems still occur. If not, plug the hub back in between the display and the keyboard, see if the problem comes back.

I'm wondering if perhaps the hub could be causing problems. Ok, the drives are not connected to it directly, but it is on the USB chain and that could be causing problems.

How many USB ports do you have on the front and rear of the machine? I know first gen G5's had two on the back and one on the front. I think newer ones are similar to the Mac Pro, with two on the front and three at the rear.

RichB

Hi Marcus,
Correct - rjb24 is me.
Yeah, id been slowly moving the disks and everything off the Trust hub anyway, in fact I lied, the keyboard was in the ACD (Apple Cinema Display). I'd already moved on from that config, anyway, so actually nothing was in the Trust hub. The Trust hub is now not connected, both disks are directly in the back of G5 and the ACD in the front or G5. Both disks are working and everything seems normal. I'll see what happens tomorrow.
This is a first gen G5 so you are correct, 2 on the back, 1 on the front.
The hub is fairly new so couldn't say if the problem happened before. I bought the 500gb with Time Machine in mind so that combination never occurred under Tiger.

Markus

Ok, keep me/us updated as to wether you see the problem with the current config. I can see why you'd go with a hub when you don't have that many ports. I've been mulling over a USB/Firewire combo hub as I'd like my three external drives, keyboard, iPod dock, media reader, printer, and other bits and pieces all connected to the iMac at the same time and I'm running out, well actually, I've run out of ports.

I *know* I can daisy chain drives but I just don't like doing that, mainly as it means to access the second and third drive I need to power on the first and/or second drives, which due to the way they are partitioned means I have far to many volumes on my desktop, and my poor eyes cannot cope

RichB

Starting to suspect Time Machine more now.
Nothing much running this morning, just copying stuff of the 250 onto the internal and 500gb to erase the 250, TM kicked in about 5 minutes ago and now its all locked up again, apps that were running, such as Safari, still work but it wont open anything new until I turn off the 250gb disk.

Going to try and turn off TM now and see if the same happens for the rest of the day.

streetbeats

Do you produce ScoobyDoo?
I run a few labels and a studio.

Havent upgraded my studio machines yet, they're running Tiger still, but my MacBook is running 10.5.1, with no problems thus far.

@RichB

I did see some weirdness with the initial Leopard release, but 10.5.1 seems to have made things far more stable, particularly where external drives are concerned, and Time Machine.
The drives I use with this machine are also USB2.
Do your drives have FW connections also?
Might be worth trying the FW bus if so and see if the problem still occurs?

Echo

ScoobyDoo only works on SSL boards but a friend of his can help you out with one

RichB

All my disks are USB2 only

Ian Cook

Well my new Macbook pro just stuffed my external USB disk up using time machine, have ordered disk warrior to try to fix it as its praises are sung on the apple forums.

I had formatted the external as HFS+ with GUID as recommended and it now says my itunes library that was on the 120gb is over 1.46tb in size ! erm i dont think so ! lol everytime i plug it in it says its unrepairable and i need to back the data up an reformat, however if i try to copy the data off it says there isnt enough disk space available because its reporting it as several TB of data ! lol

Have turned time machine off for now and will see what happens, everything else seems ok in Leopard though !

streetbeats

Any update RichB? Or indeed Ian Cook?

RichB

LOL - Strangely here is an update to the weird VM issue that I was just on my way back to report...

I found this post which is indeed true:
"When you are running the Activity Monitor and see something like 67 GB VM use, before you panic, go up to the menu and under View->Columns add a non-default item such as #Ports. You will like the results.
Mine drops from a "fictitious" 67 GB to a realistic 8 GB. "

My https drop from 6.59gb to 31.62 mb.

Machine is still locking up regularly though.
Now with 4gb of RAM

Ian Cook

This is really bizarre, my external hdd that got messed up is still knackered on the file system front but is repairable, my other external HDD is a Toshiba 320gb one, and it works perfectly on Time Machine, i cant explain this at all !

I have no slow performance issues or any other probs apart from time machine, its a 2.4ghz mac book pro with 2gb ram

Ian Cook

oh quick update fix has finished and other external is useable again now, but i wont risk it for time machine

streetbeats

weird indeed Ian - and good call on not using it for TM

Rich. Strangely enough I have had a few hangs with Leopard recently (10.5.1).
Not Kernel Panic, just a complete lock up.
However, if I have closed the lid on the MacBook and left it a few seconds, then reopened it, it has come back to life. Cant explain it, and its not always worked first time, but it has worked!

ScoobyDoo555

LOL @ Mike!!! (Echo) Gonna be down your way in the new year m8 - will take you out for a bit to eat eh?

StreetBeats - yeh I do a bit pm me - there's a few producers/engineers on this site.

Still leaving Leopard for a while - sounds as bad as (well, not quite as bad) the original OSX

Dan

RichB

One thing I did find interesting is under Energy Saver there is the option to put hard disks to sleep where possible. Mine was checked (by default) and I wondered if this might be the reason, maybe they dont wake properly if they are USB bu I disabled it and have had the sam problem again since.

RichB

Ian you do sound like you are having the same issue as me....

Ian Cook

Rich, my Tosh external disk is working fine with TM, got no idea why though, ts backing up whenever its left on, i have 7 fully working backups of my system, the drive is fine (or seems it at the moment) but the external i was using was so messed up diskwarrior took ages to fix it, but i can read and write to it fine ! Really bizarre issue, never known a computer fussy about the hard drives it writes too lol