Directoryservices 190% CPU

I had the same problem. I work a lot with MAMP. A Mac-apache-mysql-php package for local web development. On tuesday i had it three times that the directoryservice was picking on me. A reboot is the only thing to save me.

I did all the workarounds which are descript here. Since that moment i didnt have any problems. But i have been working at school the last day. I only experience the problem at school. When i changing to another 802.1X network. Maybe that information can help Apple.

This directoryservice bug is the only thing whats still bugging me. My other problem is the kernelpanic with airport. But we all now the reason for that. And 10.5.2 is maybe gonna solve that.

Sorry for my bad englische

Ok, it happened again today, Checking the Console logs for DirectoryService it says the following:
2007-12-20 11:20:28 EST - T[0xA051EF60] - DirectoryService 5.0 (v514) starting up...
2007-12-20 11:20:29 EST - T[0xB0185000] - Plugin <Cache>, Version <1.0>, processed successfully.
2007-12-20 11:20:29 EST - T[0xB0185000] - Plugin <Configure>, Version <3.0>, processed successfully.
2007-12-20 11:20:29 EST - T[0xB0185000] - Plugin <Local>, Version <1.0>, processed successfully.
2007-12-20 11:20:29 EST - T[0xB0185000] - Plugin <LDAPv3>, Version <3.0>, processed successfully.
2007-12-20 11:20:29 EST - T[0xB0185000] - Plugin <Search>, Version <3.0>, processed successfully.
2007-12-20 11:20:29 EST - T[0xB0185000] - Plugin <BSD>, Version <2.0>, processed successfully.
2007-12-20 11:20:29 EST - T[0xB030B000] - Registered node /Configure
2007-12-20 11:20:29 EST - T[0xB040F000] - Plug-in LDAPv3 state is now inactive.
2007-12-20 11:20:29 EST - T[0xB0491000] - Registered node /Search
2007-12-20 11:20:29 EST - T[0xB0491000] - Registered node /Search/Contacts
2007-12-20 11:20:29 EST - T[0xB0491000] - Registered node /Search/Network
2007-12-20 11:20:29 EST - T[0xB0491000] - Plug-in Search state is now active.
2007-12-20 11:20:29 EST - T[0xB030B000] - Plug-in Configure state is now active.
2007-12-20 11:20:29 EST - T[0xB0513000] - Registered Locally Hosted Node /BSD/local
2007-12-20 11:20:29 EST - T[0xB0513000] - Registered node /BSD/local
2007-12-20 11:20:29 EST - T[0xB0513000] - Plug-in BSD state is now active.
2007-12-20 11:20:29 EST - T[0xB038D000] - Registered Locally Hosted Node /Local/Default
2007-12-20 11:20:29 EST - T[0xB038D000] - Registered node /Local/Default
2007-12-20 11:20:29 EST - T[0xB038D000] - Plug-in Local state is now active.
2007-12-20 11:20:29 EST - T[0xB0289000] - Registered node /Cache
2007-12-20 11:20:29 EST - T[0xB0289000] - Plug-in Cache state is now active.
2007-12-20 11:20:29 EST - T[0xB0185000] - Plugin "Active Directory", Version "1.6", is set to load lazily.
2007-12-20 11:20:29 EST - T[0xB0185000] - Plugin "PasswordServer", Version "4.0", is set to load lazily.
2007-12-20 11:20:37 EST - T[0xB0081000] - Network transition occurred.
2007-12-20 11:20:37 EST - T[0xB0081000] - Network transition occurred.
2007-12-20 11:20:38 EST - T[0xB0081000] - Network transition occurred.
2007-12-20 11:20:38 EST - T[0xB0081000] - Network transition occurred."

At this point the fan is roaring away and the Mac locks up, cant load Activiy Monitor, nothing.

I have experienced similar symptoms on my MBP without Parallels Desktop or EarthDesk installed.

This problem, on my platform, occurs intermittently rather than being reproducible on demand.

If it were a Leopard "bug," then it should be reproducible on demand, all users should be experiencing it, and it should happen on the floor model Macs at all Apple Stores. Further, if this were the case, it most likely would receive huge media attention.

I think, instead, that something else is/are the culprit/s.

One possiblity is a hardware problem (e.g., Airport card; RAM modules) which could account for the evidently limited scope of affected users but does not account for the intermittent nature of the problem

Regardless, to check this out, I have given my MBP to my local Apple Store for evaluation.

One other possibility is that the culprit in this scenario is third-party software which needs to be updated for Leopard compatibility. Again, however, if this were the case, the symptoms should be happening more reproducibly rather than intermittently.

If anyone in this thread is able to reproduce the symptoms on demand, could you create a new user account (or use the Guest one in Leopard) and perform the exact same steps that previously have produced the symptoms?

Interesting suggestions - and you're right that this is an intermittent rather than a universal, infinitely repeatable bug. However, I think there is already a reasonable amount of evidence in this thread to suggest that it is still a bug associated with Leopard, although it appears to be a bug that is only associated with handling a particular kind of situation. Such a bug would not necessarily be experienced by the majority of users, nor in Apple stores. It seems to be associated with changing or dropping network connections (e.g. a laptop moving between airport networks, or dropping off one while still running), and the majority of Macs (certainly all those in Apple stores) spend their lives connected to one, reliable network connection. Also, since the consequences are merely annoying rather than destructive, media attention is likely to be non-existent for a bug affecting only a minority of users.

Although none of the suggestions here got rid of the bug entirely for me (changing settings in Directory Utility and/or repairing permissions), I have not seen it since I changed the network setup from an airport network with a very weak signal to using a wired ethernet connection bridged through an old iBook... circumstantial evidence, but evidence which supports the "dropping network connection" theory.

I will certainly be interested to see what (if anything) the Apple store can find with your MBP - I'm guessing they'll have trouble reproducing the bug, but I'm always happy to be corrected!

Happy New Year all. I suspect this topic won't disappear in 2008, so to add to the many voices above I experience this problem sometimes when offline - not so good when showing photos to elderly relatives over Christmas. The rising fan usually gives me enough notice to save and close programmes before restarting the machine in the normal fashion, otherwise it's a forced restart.

It's when I'm offline, don't have my Time Machine disk attached and I'm not doing anything special. Often I can't open Activity Monitor, but I'm tempted to try force quitting the process if I get a chance next time. (I planned on doing a clean install to celebrate the new year in case that fixed it, but if 10.5.2's only a fortnight away I'll live with it til then and hope there's a fix a'coming.)

To add to previous comment, Apple support (from the UK) say they're aware of this problem - once the operator's checked with a higher level of technician - and they suggest backing up and running a clean install from the Leopard installation disc.

I'm still going to give 10.5.2 a chance, but I foresee a day of rebuilding before the end of January!

Hope this helps, David.

I was having constant issues (and I mean all the time) with directoryservice although my cpu use never went over 100%. For other reasons I did reinstall my OS (archive & install) and it is all gone now, already for a while. It is nasty but it does seem to work.

P.

Hi folks,
I have the same problem from time to time, and I think I have recognized what's causing it (at least on my MacBook Pro)

Basically sometimes I launch iTunes (which in turn launches last.fm application to scrobble what I listen to and send the info to last.fm website) with the computer connected (via Ethernet) to the internet... I start playing music then I disconnect all the cables and move the laptop away
Then after a while I discover DirectoryServices stealing loads of CPU (and therefore making the laptop boiling).

I agree with the people connecting this problem to the lack of a network connection, since last.fm tries to send info to the last.fm website, but of course cannot since there's no network connection... (I do this because when I reconnect to the network the app sends all the information about the songs I listened to while disconnected)

But the thing is that once the network connection is re-established, DirectoryServices stays up at 100% (or more) CPU, therefore I need to reboot (so annoying)

It happens to me from time to time, and I always noticed it in the situation described above...

I hope this will be fixed soon!
Tom.

mac os 10.5.1, powerbook 12", 1.5Ghz, 1.25Gb ram, all updates.

I can duplicate this problem consistantly. Whatever application is using the most CPU, DirectoryService will mirror that app's usage, every single time.

Mail, Safari, Finder, whatever is using the most will be mirrored, usually within 2% of usage.

This is pretty annoying, considering it slows the machine by 1/2, and making attempting anything CPU intensive unusable.

I want my fantastic G4 back, apple, not this P3 500Mhz that DirectoryService is slaying. 😉

thanks guys for getting on top of it.

I have been having trouble with DirectoryService locking up causing my entire system to go down. As soon as it is locked up, DirectoryService takes up 190% CPU time, and all programs fail to launch since they are trying to do something network related, or looking up of usernames even since that is all run through a single daemon.

The following messages can be found in /var/log/system.log

Jan 6 19:51:53 WideLoad DirectoryService[637]: DSMutexSemaphore::SignalDebug - Error 1 for mutex 0x0020D470 - Caller CSearchPlugin.cpp:2629
Jan 6 19:51:53 WideLoad DirectoryService[637]: If you run under the debugger, it will automatically break here.
Jan 6 14:51:57 WideLoad DirectoryService[637]: DSMutexSemaphore::SignalDebug - Error 1 for mutex 0x00216814 - Caller CDSLocalPluginNode.cpp:503
Jan 6 14:51:57 WideLoad DirectoryService[637]: If you run under the debugger, it will automatically break here.
Jan 6 14:51:57 WideLoad DirectoryService[637]: DSMutexSemaphore::SignalDebug - Error 1 for mutex 0x0020D470 - Caller CSearchPlugin.cpp:2629
Jan 6 14:51:57 WideLoad DirectoryService[637]: If you run under the debugger, it will automatically break here.

Jan 6 16:26:15 WideLoad DirectoryService[32]: DirectoryService(32,0xb0524000) malloc: * error for object 0x2278a0: incorrect checksum for freed object - object was probably modified after being freed.\n * set a breakpoint in malloc errorbreak to debug
Jan 6 16:26:15 WideLoad DirectoryService[32]: DirectoryService(32,0xb0524000) malloc: * error for object 0x2278e0: incorrect checksum for freed object - object was probably modified after being freed.\n * set a breakpoint in malloc errorbreak to debug
Jan 6 16:26:15 WideLoad com.apple.DirectoryServices[32]: DirectoryService(32,0xb0524000) malloc: * error for object 0x2278a0: incorrect checksum for freed object - object was probably modified after being freed.
Jan 6 16:26:15 WideLoad com.apple.DirectoryServices[32]: * set a breakpoint in malloc errorbreak to debug
Jan 6 16:26:15 WideLoad com.apple.DirectoryServices[32]: DirectoryService(32,0xb0524000) malloc: * error for object 0x2278e0: incorrect checksum for freed object - object was probably modified after being freed.
Jan 6 16:26:15 WideLoad com.apple.DirectoryServices[32]: * set a breakpoint in malloc errorbreak to debug

Jan 14 00:51:54 wideload DirectoryService[32]: DSMutexSemaphore::SignalDebug - Error 1 for mutex 0x00215AE4 - Caller CDSLocalPluginNode.cpp:1079
Jan 14 00:51:54 wideload DirectoryService[32]: If you run under the debugger, it will automatically break here.
Jan 14 00:51:54 wideload DirectoryService[32]: DSMutexSemaphore::SignalDebug - Error 1 for mutex 0x00209800 - Caller CDSLocalPlugin.cpp:3385
Jan 14 00:51:54 wideload DirectoryService[32]: If you run under the debugger, it will automatically break here.
Jan 14 00:52:01 wideload DirectoryService[32]: DSMutexSemaphore::SignalDebug - Error 1 for mutex 0x00215240 - Caller CSearchPlugin.cpp:2629
Jan 14 00:52:01 wideload DirectoryService[32]: If you run under the debugger, it will automatically break here.
Jan 14 00:52:01 wideload DirectoryService[32]: DSMutexSemaphore::SignalDebug - Error 1 for mutex 0x00215240 - Caller CSearchPlugin.cpp:2629
Jan 14 00:52:01 wideload DirectoryService[32]: If you run under the debugger, it will automatically break here.

I started getting more that 80% CPU (mind you, my old G4 has only 1 CPU :)) eaten up by DirectoryService recently.
It must be some buggy code that has gotten into MacOSX, accross the platforms, in recent updates, 10.4.11 (or one or two updates before this) in my case.

If you'll notice the thread starters posted comments about this problem over 2 1/2 months ago ! On apple's own discussion threads, not some out-of-the-way forum.

Apple, why hasn't this been answered and resolved already ?

Shame shame.

i completely agree... I am facing the same problem and force-rebooting my laptop everytime like an old windows box is shameful.

I'm right behind you; it's taking too long for someone to categorically work out what the problem is let alone for Apple to find a fix.

On Apple (UK) support's advice I spent a weekend doing a clean install, reinstalling everything from scratch with new downloads, etc. The problem's still there from time to time, although I now try not run programmes that require the net if I'm not properly connected, which helps. Following on from one of the comments above I may give up scrobbling for a while (I use iScrobbler).

Come on Apple! Something for this and improved Airport Extreme disk mounting in 10.5.2 please.

David.

Yet another person here who is experiencing a similar problem.

Apple - please fix it ASAP