Looks like no one’s replied in a while. To start the conversation again, simply ask a new question.

Level 1

145 points

Directory Service Crashing

I have been testing leopard server and have run into a problem that puzzles me. I hope that somebody can provide some insights on this issue.

Installed 10.5 server, advanced with mobile users and time machine backups. Upgraded mobile accounts from Tiger were less than desirable (sync issues). 10.5 users were fine except that the directory service crashed consistently during home directory syncing.
Did a fresh re-install with upgrade to 10.5.1 before configuration; still the same.

Did even a standard install with standard users only (basic services to make sure that no manual DNS or OD config. caused the problem ) and the Directory Services still go out on me.

This is running on a Xserve G5 dual with 6gig of RAM & 500gb raid1 mirror. That machine ran 10.4 perfectly for the last year plus.

Below are the shortened results form the crash reporter. Any ideas? I was really hoping that I can use leopard server to use iCal server & time machine.

Thanks!

Process: DirectoryService [26]
Path: /usr/sbin/DirectoryService
Identifier: DirectoryService
Version: ??? (???)
Code Type: PPC (Native)
Parent Process: launchd [1]

Date/Time: 2007-11-25 20:33:56.179 -0600
OS Version: Mac OS X Server 10.5.1 (9B18)
Report Version: 6

Exception Type: EXC BADACCESS (SIGSEGV)
Exception Codes: KERN INVALIDADDRESS at 0x000000004b009698
Crashed Thread: 2

Thread 0:
0 libSystem.B.dylib 0x93f689d8 mach msgtrap + 8
1 libSystem.B.dylib 0x93f6f8fc mach_msg + 56
2 com.apple.CoreFoundation 0x922f9664 CFRunLoopRunSpecific + 1828
3 com.apple.CoreFoundation 0x922f9ef4 CFRunLoopRun + 60
4 DirectoryService 0x000156e4 main + 2356
5 DirectoryService 0x00014d78 start + 64

Thread 1:
0 libSystem.B.dylib 0x93f689d8 mach msgtrap + 8
1 libSystem.B.dylib 0x93f6f8fc mach_msg + 56
2 com.apple.CoreFoundation 0x922f9664 CFRunLoopRunSpecific + 1828
3 com.apple.CoreFoundation 0x922f9ef4 CFRunLoopRun + 60
4 DirectoryService 0x00020eb0 CPluginRunLoopThread::ThreadMain() + 228
5 ...ectoryServiceCore.Framework 0x0015a09c DSCThread::Run() + 52
6 ...ectoryServiceCore.Framework 0x0015a2f4 DSLThread::_RunWrapper(void*) + 96
7 libSystem.B.dylib 0x93faabf8 pthreadstart + 316

Thread 2 Crashed:
0 libsqlite3.0.dylib 0x94ae6850 sqlite3 resultvalue + 33024
1 libsqlite3.0.dylib 0x94ae5728 sqlite3 resultvalue + 28632
2 libsqlite3.0.dylib 0x94af1e5c sqlite3 resultvalue + 79628
3 libsqlite3.0.dylib 0x94af1db4 sqlite3 resultvalue + 79460
4 libsqlite3.0.dylib 0x94b3c3dc sqlite3 gettable + 31548
5 libsqlite3.0.dylib 0x94b3fbf4 sqlite3 gettable + 45908
6 libsqlite3.0.dylib 0x94b3fe64 sqlite3_step + 148
7 DirectoryService 0x0006d31c CDSLocalPluginNode::sqlExecSync(char const*, unsigned long) + 100
8 DirectoryService 0x0006eda4 CDSLocalPluginNode::AddRecordIndex(char const*, char const*) + 512
9 DirectoryService 0x0006f9b0 CDSLocalPluginNode::FlushRecord(__CFString const*, __CFString const*, __CFDictionary const*) + 1308
10 DirectoryService 0x0006972c CDSLocalPlugin::CloseRecord(sCloseRecord*) + 304
11 DirectoryService 0x000a4d80 BaseDirectoryPlugin::ProcessRequest(void*) + 520
12 DirectoryService 0x00069564 CDSLocalPlugin::ProcessRequest(void*) + 600
13 DirectoryService 0x00012874 CHandlerThread::RefDeallocProc(unsigned long, unsigned long, CServerPlugin*) + 240
14 DirectoryService 0x0000a73c CRefTable::RemoveRef(unsigned long, unsigned long, long, unsigned long, bool) + 880
15 DirectoryService 0x0001d6fc CRequestHandler::DoCloseRecord(sComData*, long*) + 220
16 DirectoryService 0x00009ce4 CRequestHandler::GetRequestData(sComData*, long*, bool*) + 808
17 DirectoryService 0x00002360 CRequestHandler::HandlePluginCall(sComData**) + 192
18 DirectoryService 0x00003068 CRequestHandler::HandleRequest(sComData**) + 76
19 DirectoryService 0x0002d4b8 dsmig do_apicall + 380
20 DirectoryService 0x0005de5c Xapicall + 452
21 DirectoryService 0x0005da80 DirectoryServiceMIG_server + 164
22 DirectoryService 0x00024330 dsmig demux_notify(mach_msg_headert*, mach msg_headert*) + 96
23 libSystem.B.dylib 0x93ff209c mach msgserver + 364
24 DirectoryService 0x00021130 CMigHandlerThread::ThreadMain() + 312
25 ...ectoryServiceCore.Framework 0x0015a09c DSCThread::Run() + 52
26 ...ectoryServiceCore.Framework 0x0015a2f4 DSLThread::_RunWrapper(void*) + 96
27 libSystem.B.dylib 0x93faabf8 pthreadstart + 316

Xserve's, misc. Mac models, Mac OS X (10.4.10)

Posted on Nov 25, 2007 6:58 PM

74 replies

BASD IT

Level 1

0 points

Feb 28, 2008 6:18 AM in response to Bradford Schwie

Oh, sorry. You made it sound like you "got Leopard Server" as if you already had Tiger and just bought 10.5.

I don't blame you for being wary, but something is better than nothing. TMB works via a Leopard Client AFP share too (it looks like you already knew this). I don't know if this will help you or not, but here's a possible setup.

2 Leopard clients, both have a partition for TMB. Partition shared on both. Run the command I recommended. Find each other. Viola.

I don't know your setup, but this DirectoryService crashing issue is supposed to only happen when you have OD and AFP running on the same server. Maybe you could use Leopard Client on someone's mac to share a backup volume at the office somewhere and authenticate to the Leopard Server's OD, turning off AFP on the server, of course.

Maybe you could just run your Leopard server in "Stand Alone" mode?

There are a few possibilities. Just trying to be helpful. We're all hoping and waiting for a solution to this frustrating issue.

P.S. Maybe try this hint?
http://www.macosxhints.com/article.php?story=20071108020121567

Mar 6, 2008 10:54 AM in response to BASD IT

My users are network users. When directory service crashes, the users who were lucky enough to already have been logged in can continue working, but they have lost access to folders/databases on the afp volume. When the cron job kicks in, those folders/databases may or may not be there for them again. I think we're going back to 10.4.11 until Apple works this out. I don't have time to cope with the animosity and frustration being experienced by our users.

I wanted to "harness the power of Leopard to run my network", but really need the stability of Tiger.

Mar 6, 2008 1:57 PM in response to Alexander Hosmer

I feel similar to you. I bought Leopard Server as our first Mac OS X Server license for AFP with other services on the same box (Mac Pro). Now that I can't use everything simultaneously, I feel like I should be offered Tiger Server until they figure it out.

Mar 27, 2008 4:37 AM in response to Wolfman

I made at a customer a new fresh clean install of mac os x 10.5.2 server on a xserve G5, DirectoryService is crashing every now and then causing AFP not to accept logins. By going to server admin and changing Authentication method to something else ( Any Method, -> Kerberos -> Any Method, SAVE ) and back, then press save, it works again. I've tested this approach several times now at the customer and those still logged in didn't loose their connectivity. No restarts required.

Process: DirectoryService [48935]
Path: /usr/sbin/DirectoryService
Identifier: DirectoryService
Version: ??? (???)
Code Type: PPC (Native)
Parent Process: launchd [1]

Date/Time: 2008-03-27 08:47:28.129 +0200
OS Version: Mac OS X Server 10.5.2 (9C31)
Report Version: 6

Exception Type: EXC BADACCESS (SIGSEGV)
Exception Codes: KERN INVALIDADDRESS at 0x000000004b005a08
Crashed Thread: 2

Thread 0:
0 libSystem.B.dylib 0x95fe1978 mach msgtrap + 8
1 libSystem.B.dylib 0x95fe889c mach_msg + 56
2 com.apple.CoreFoundation 0x94810618 CFRunLoopRunSpecific + 1828
3 com.apple.CoreFoundation 0x94810ea8 CFRunLoopRun + 60
4 DirectoryService 0x00015c50 main + 2340
5 DirectoryService 0x000152f4 start + 64

Thread 1:
...

-rw-r----- 1 root admin 15617 Mar 19 10:26 DirectoryService 2008-03-19-102644localhost.crash
-rw-r----- 1 root admin 15986 Mar 19 12:27 DirectoryService 2008-03-19-122723localhost.crash
-rw-r----- 1 root admin 13862 Mar 26 08:30 DirectoryService 2008-03-26-083034localhost.crash
-rw-r----- 1 root admin 15928 Mar 27 08:47 DirectoryService 2008-03-27-084728localhost.crash

Apr 1, 2008 2:56 AM in response to Wolfman

This seems to be exactly the issue we're having.

We have a brand new Mac Pro, with Leopard server. Our first OS X server, and currently it is serving only 3 users during testing (only a small organisation of 7 anyway).

So, it all seemed to be fine, and then I got reports from users that Time Machine backups weren't working. They also reported they couldn't browse the machine from the Finder. Yet web site authorisation worked, so it seemed Open Directory was working.

Rebooted today, and they said it was all working. And then later that day it all happened again.

Finding this discussion, I checked and lo and behold:

yamato% ls
DirectoryService 2008-03-15-152920localhost.crash
DirectoryService 2008-03-25-114211localhost.crash
DirectoryService 2008-04-01-113348localhost.crash
DirectoryService 2008-04-01-113350localhost.crash
DirectoryService 2008-04-01-113504localhost.crash
DirectoryService 2008-04-01-122730localhost.crash

My worry here is that we are now at 10.5.2, and still no fix - but it looks like people were reporting this to Apple from 10.5.0. What are our chances of a fix in 10.5.3? Leopard server seems to be so buggy - what is going on here?

Apr 3, 2008 2:08 AM in response to Wolfman

Does anyone have a cron job that has worked? I installed the cron job as discussed here, as root, but it made no difference when, once again, DirectoryService died today. The only solution that worked was when I manually stopped and started AFS from the Server Admin GUI.

Maybe the HUP signal isn't doing the job, but we can't have it kill the process and restart it every 2 minutes, as I would expect that would cause issues if someone was in the middle of a Time Machine backup.

I suppose a script could be created that watched the CrashReporter logs and noticed when Directory Service died, and then killed off and restarted AFS. But that seems extreme...

Pagrash

Level 1

84 points

Apr 4, 2008 10:46 AM in response to Phil Houtz

Don’t know if this helps, but I had the same problem after a clean install of 10.5.0 on a new Mac Pro.

The only clues were that it was crashing once an hour when it did a checkpoint save, and, the Disk Utility was showing lots of disk permission problems.

The problem was because disk utility on 10.5.0 could not partition RAID (mirror) drives properly, but would allow you continue the installation.

Eventually, I used a Tiger disk to partition the drives before installing Leopard, and the crashes stopped.

Apr 4, 2008 6:46 PM in response to Wolfman

Ok, the cron job mentioned here seems to be wrong. I believe it should be:

*/2 * * * * killall -HUP AppleFileServer

installed as root. Waiting for the crash to occur again so I can confirm it works...

Apr 7, 2008 6:20 AM in response to BASD IT

Same problem here. OD has been locking up with more frequency now, first every 2 weeks then one, now its every 3 or 4 days. I've tried turning off services including AFP but it still crashes. When it crashes the server is toast, no mail, no ard, reboot is the only option.

I had to configure LOM just to reboot the machine from remote.

One new problem has cropped up... Xserve no longer safe boots, just restarts over and over when holding down shift. I really don't want to have to go through a rebuild again.

Apple! What's going on? Please help!

skp574

Level 1

0 points

Apr 8, 2008 2:48 AM in response to Jamie Curmi

Would someone be willing to do a step by step guide to set up the cron job. I downloaded Cronnix to help do it for me, but a day later the server locked out again which makes me think it didn't work. From reading the above posts I don't know is people are experiencing exactly the same issue.

If I look in /etc/ I can't see a cron entry.

I find with our system after a totally random time (6 hours to 3 days) the server fails to authenticate anyone in the OD and even the root local user account on the server. The Server Admin GUI can no longer monitor the running services and Workgroup Manager can't do anything either. I can't authenticate and use the 'su' command in a terminal window so the only thing I can do it to restart the server.

I'm running 10.5.2 on a MacPro 2 x 2.0 GHz.

This is the bane of my life at the moment. Any help or guidance will be greatly appreciated.

Apr 8, 2008 2:59 AM in response to skp574

The cron job doesn't seem to work anyway, as it turns out (from my testing). But what you are experiencing seems more extreme than I am seeing.

What I see (and have reported to Apple) is DirectoryService crashing 2-3 times a day. When it does, it restarts, but AFP stops authenticating. But I can still ssh to the machine etc.

The only solution I have come up with that works is that I have a launchd file that runs a script if something crashes, and the script checks to see if the crash was DirectoryService, and if so, restarts AFP.

This works - except that if someone is connected to the server (say doing a backup), it bumps them and they get a message saying they have been disconnected. That isn't pleasant, and gives the appearance that the server keeps crashing daily. But so far it is the only thing I've found that works.

Can share it with anyone who wants it. I'm trying to find a solution that works without having to restart, but nothing else seems to work consistently that I've tried.

Ultimately though, we need a fix for this from Apple ASAP. This bug is devastating for many users out there - and is embarrassing as **** for those of us who've worked to convince companies to try and use OS X Server rather than Linux or Windows.

skp574

Level 1

0 points

Apr 8, 2008 4:13 AM in response to Jamie Curmi

Hi Jamie,

Thanks for your reply. Looks like, until there is a fix, I may have to take a more drastic approach and set up a cron job to restart the server over night. I am relatively lucky that our server is under a light load, so the OD crash/lock up happens mostly every couple of days.

Apr 8, 2008 5:04 AM in response to skp574

It is interesting that we have only 3 users of the server at present, so we aren't stressing it at all. Yet DirectoryService crashes 2-3 times a day.

skp574, do you actually see the DirectoryService crashes in /Library/Log/Crashes on the server, and can confirm that your loss of authorisation occurs soon after these crashes?

Apr 8, 2008 6:26 AM in response to Wolfman

Hi All.

I've created a launchd daemon that restarts AFP if DirectoryService crashes, which seems to work (though isn't without issues of its own). For those who've tried everything and nothing works, you may want to give it a try. May at least be a work around until Apple fix this (please Apple, fix this soon), or a tool you can play with to experiment with your own fixes.

http://curmi.com/blog/2008/04/08/leopard-server-and-directoryservice-crashes/

Full instructions on how to install and uninstall are included.

Hopefully it is useful to someone.

Jamie

Apr 9, 2008 11:49 PM in response to Jamie Curmi

You didn't try my approach by simply going to server admin, afp, change auth method from any to something else and back to any?

(you could script serveradmin settings afp:authenticationMode = "standard andkerberos" )

Directory Service Crashing