Apple Support Communities > Servers and Enterprise Software > Mac OS X Server v10.4 and earlier > Discussions
This discussion is archived
6871 Views 61 Replies Latest reply: Apr 19, 2007 6:20 AM by Justin Hill4
Currently Being ModeratedJan 29, 2007 2:02 AM (in response to haddock92)This is a known problem at the moment and only affects the Universal version of Mac OSX Server. It is only triggered when you do a find either on the server itself or via a client. The wired memory will increase until finally loacking up the server. Apple is working on the problem which will hopefully be solved in the next update to the operating system or via a seperate patch.
Hope this helps everyone, we only discovered it when we installed the universal version on a PPC machine, we rebuilt the machine and everything was fine after opening a ticket with Apple and discussing the problems.
Currently Being ModeratedJan 29, 2007 4:41 AM (in response to Craig Sayer)
Hope this helps everyone, we only discovered it when
we installed the universal version on a PPC machine,
we rebuilt the machine and everything was fine after
opening a ticket with Apple and discussing the
How did you fix the problem - is there a non-Universal/PPC-only installer or set of disks?
PowerMac Mac OS X (10.4.8) Dual 2.5GHz G5 2.5G RAM
Currently Being ModeratedJan 31, 2007 10:08 AM (in response to Craig Sayer)New theory for 10.4 Universal on PPC hardware: acls are the culprit.
After talking to some other sysadmins about their experience with 10.4 Universal it turned out that only those had this problem who where using acls.
Currently Being ModeratedJan 31, 2007 10:15 AM (in response to slowfranklin)ACL's off on all volumes. Not even available for use. The PPC process and Search issue both happen on my Mac Pro w/Universal 10.4.8 installed. Services are AFP, SMB, Web, NTP, Mail, SU, OD. I will test the search via an SMB Mount later.Mac Pro 2 x 2.66Ghz intel 2006, Mac OS X (10.4.8), ATTO SCSI320 2CH, Highpoint 2322 SATA 2 RAID
Currently Being ModeratedFeb 7, 2007 4:27 PM (in response to Michael Ojaste)Thanks everyone for doing my research for me on the memory leak! My experiences directly reflect those posted. I am discouraging searches and doing daily restarts until a fix is available.
Please someone post when there is a fix available.
thanks, DaveMacBook Pro, Mac OS X (10.4.8)
Currently Being ModeratedFeb 7, 2007 11:04 PM (in response to thufir)Is everybody reporting to http//radar.apple.com/ ? If not, please do so.
Apple is quite responsive this time on the bug report I filed.
Thinks to include:
your observation of
$ top -l 0 | grep kernel_task
and the output of
Currently Being ModeratedFeb 13, 2007 6:53 PM (in response to slowfranklin)I reported this problem to Apple in december. They are supposedly working on it.
BTW, to stop the controversy with Rosetta. I reproduced the problem in like 15 mins with the machine, freshly rebooted with no retropect daemon running.
I fed Apple with
Screenshot of activity monitor before it happens
Screenshot of activity monitor while it happens
Countinuous top output (fresh from a reboot) from a "Clean state" to a "Problem state"
Countinuous AFP output (fresh from a reboot) from a "Clean state" to a "Problem state" with one user doing search.
Output stopped at one point, because the server crashed.
From an enterprise POV, the service is not fast enough.
Currently Being ModeratedFeb 14, 2007 12:26 PM (in response to Moofo)Ditto here. I also reported way back in October and Apple's engineering is apparently still working on this.
I wonder if I should be hopeful that OS 10.4.9 will have a fix ...2.66 ghz Mac Pro, 2 GB RAM, Mac OS X (10.4.8), running OS X Server
I'm seeing massive 10.4.8 memory leaks on a G4 Mac mini running 10.4.8 Server, but only when iTunes is left running. Over a couple of days, I end up with 10+ swap files, even if iTunes is never played or searched. Without running iTunes, I never even get to two swap files over a period of weeks. This immediately started happening after the update to 10.4.8.
Hope they fix it...Quad G5; 3 x mini; AI PB; PB180; iPod; Shuffle,and much, much more, Mac OS X (10.4.8)
I'm sorry but why in gawd's name are you running iTunes on a "Server" ????
"Yah, my server really lags when I'm recording HD video and burning a CD and playing UT in windowed mode"
Unrealistic expectations typically yield unhappy results.
Never mind that the drive in a Mini isn't rated for continuous use and will do good deal to increase your overall latency.
10.4.7 server did not memory leak, whereas 10.4.8 does. That's all I am saying. Don't come at me with your fancy-dan stick when you know essentially nothing about my particular situation.
Don't shoot the messenger.Quad G5; 3 x mini; AI PB; PB180; iPod; Shuffle,and much, much more, Mac OS X (10.4.8)
No idea what you're referring to, especially here in this context.
If you have verifiable memory leaks under 10.4.8 where you did not under 10.4.7, then by all means say as much.
If you expect proper performance from OS X Server on what is undisputably Apple's most client-style box while running client-type apps, then may t3h f0rce be with you.
If you have verifiable memory leaks under 10.4.8
where you did not under 10.4.7, then by all means say
I'm not exactly sure how else one might interpret this:
"This immediately started happening after the update to 10.4.8."
And this isn't about "proper performance", I simply chimed in on an existing thread to add information. Discussion was suggesting that this was only being seen with Rosetta apps on Intel boxes. My post was to point out that the problem may be other than that, and that we all keenly await 10.4.9 Server.Quad G5; 3 x mini; AI PB; PB180; iPod; Shuffle,and much, much more, Mac OS X (10.4.8)
Over a couple of days, I end up with
10+ swap files, even if iTunes is never played or
searched. Without running iTunes, I never even get
to two swap files over a period of weeks.
To get to the point (and not to say we are not looking for others to report) we are talking here about REAL memory that is being eaten up by the kernel_task and not being properly released back to the system under 10.4.8. Your talking about Swap or VM. Apples and Oranges in my book. If you where to tell me that upon running iTunes on 10.4.8 server the kernel_task Real Memory shot up and was not being given back to the system well then your on the right thread. Your problem is on PCC and I'd venture to say unrelated to the problem we have here which deals in real not swap memory.Mac Pro 2 x 2.66Ghz intel 2006, Mac OS X (10.4.8), ATTO SCSI320 2CH, Highpoint 2322 SATA 2 RAID