- launchd-dev - lists.macosforge.org

SMJobBless() fails to replace outdated version
by James Foster 27 Jan '17

27 Jan '17

I am attempting to replace a helper tool. Most of the time if I increment the version (of my application and of the helper tool) then the new tool will be installed. Sometimes it won't be installed and I have to delete the old one manually. The problem seems to occur about every dozen versions or so, and once it happens the replacement will not occur even if I change the version numbers again. https://developer.apple.com/reference/servicemanagement/1431078-smjobbless?… <https://developer.apple.com/reference/servicemanagement/1431078-smjobbless?…> states "If the job is already installed, success is returned.” How does SMJobBless() determine that the job is already installed? What do I need to do, beyond creating an entirely new job name, to make it look different enough to replace? Following is my code and console output (from macOS 10.12.2). James Foster CODE: - (void)installHelperTool; { OSStatus status = AuthorizationCreate(...); // handle errors NSLog(@"existing: %@", [self system:@"/usr/bin/otool -P " "/Library/PrivilegedHelperTools/com.GemTalk.GemStone.Helper | " "grep -A 1 VersionString | tail -1"]); NSLog(@"replacement: %@", [self system:@"/usr/bin/otool -P " "GemStone.app/Contents/Library/LaunchServices/com.GemTalk.GemStone.Helper | " "grep -A 1 VersionString | tail -1"]); CFErrorRef cfError = nil; if (SMJobBless(kSMDomainSystemLaunchd, (__bridge CFStringRef)@kHelperIdentifier, authRef, &cfError)) { NSLog(@"SMJobBless(%s,...) returned true; attempting (%s).", kHelperIdentifier, kShortVersionString); NSLog(@"now found: %@", [self system:@"/usr/bin/otool -P " "/Library/PrivilegedHelperTools/com.GemTalk.GemStone.Helper | " "grep -A 1 VersionString | tail -1"]); } else { AppError(@"Helper tool installation failed: %@", [(__bridge NSError*) cfError localizedDescription]); } } CONSOLE: default 14:25:04.368568 -0800 authd Succeeded authorizing right 'com.apple.ServiceManagement.blesshelper' by client '/Users/jfoster/Library/Developer/Xcode/DerivedData/GemStone-akfvfyaoddyaavgbcxnrdhnovoal/Build/Products/Debug/GemStone.app' [5626] for authorization created by '/Users/jfoster/Library/Developer/Xcode/DerivedData/GemStone-akfvfyaoddyaavgbcxnrdhnovoal/Build/Products/Debug/GemStone.app' [5626] (13,0) default 14:25:04.889238 -0800 kernel AMFI: allowing exception handler for 'GemStone' (5626) because the process is not restricted. default 14:25:04.385444 -0800 GemStone existing: <string>1.3.3.9</string> default 14:25:04.900905 -0800 kernel AMFI: allowing exception handler for 'GemStone' (5626) because the process is not restricted. default 14:25:04.397113 -0800 GemStone replacement: <string>1.4.0</string> default 14:25:04.398426 -0800 authd Succeeded authorizing right 'com.apple.ServiceManagement.blesshelper' by client '/usr/libexec/smd' [251] for authorization created by '/Users/jfoster/Library/Developer/Xcode/DerivedData/GemStone-akfvfyaoddyaavgbcxnrdhnovoal/Build/Products/Debug/GemStone.app' [5626] (3,0) default 14:25:04.408694 -0800 GemStone SMJobBless(com.GemTalk.GemStone.Helper,...) returned true; attempting (1.4.0). default 14:25:04.923798 -0800 kernel AMFI: allowing exception handler for 'GemStone' (5626) because the process is not restricted. default 14:25:04.419988 -0800 GemStone now found: <string>1.3.3.9</string>

3 5

Penalty box property?
by Nick Zolotko 28 Oct '16

28 Oct '16

Hello All, I have a launchd system service, that will stop running at seemingly odd times. The service just launches a bash script to rsync files over the network. However the launchd service will stop launching and gives me different error codes. However I can’t find any information on the launchd property of "penalty box”. Does anyone know what that does or means? When the service fails it always seems to have the state of spawning... Stderr does not produce anything. The script lives on the network and will run successfully most of the time. Once the launchd gets in this state it just fails unless I bootout/bootstrap the plist again. Any suggestions? Here is some more information. root# launchctl print system/com.laika.localsync com.laika.localsync = { active count = 0 path = /Library/LaunchDaemons/com.laika.localsync.plist state = spawning minimum runtime = 10 exit timeout = 5 nice = 20 runs = 1 successive crashes = 0 excessive crashing = 0 last exit code = 78: EX_CONFIG properties = { penalty box = 1 } root# launchctl list com.laika.localsync "LastExitStatus" = 19968; Nick Zolotko Desktop System Administrator

2 1

Re: [launchd-dev] Replacements for deprecated SMJobSubmit and SMJobRemove
by Graham Miln 21 Oct '16

21 Oct '16

Thank you Joe. I appreciate knowing to avoid shelling out to launchctl. Kind regards, Graham

1 0

Re: [launchd-dev] Replacements for deprecated SMJobSubmit and SMJobRemove for plugins
by Dave MacLachlan 19 Oct '16

19 Oct '16

Change of subject so I don't threadjack, but related... Is there a good way for a plugin to start up an xpc job? Before you say "why isn't your plugin an xpc job", it's something I want to launch from a XCTest bundle to monitor my test externally. Cheers, Dave On Tue, Oct 18, 2016 at 10:26 AM, Joe Auricchio <jauricchio(a)apple.com> wrote: > Hi Graham, > > Please *do not* shell out to launchctl. It's better to keep using the > deprecated SM functions than to run launchctl. Sorry, but replacements are > not yet available. > > There isn't presently API for a system-wide LaunchDaemon to open a Mach > IPC or XPC connection to a LaunchAgent, which I'd guess is closer to what > you'd really like to do? This would be a good enhancement request, please > file one! Radars from developers are always good ammunition for trying to > schedule our development time. (Coincidentally, now is a very good time for > this particular API request.) > > Another approach is for the LaunchAgents to add a LaunchEvent (man > xpc_events) for a notifyd (man notify_post) notification. When the daemon > wishes for an agent to start, it will post a notification; the LaunchEvent > will start agents (in all the running graphical user sessions) in response > to the notification; the agents will all check in with the daemon over MIG > or XPC when their xpc_event handler fires; and the daemon will ask the > appropriate agent to launch a process in the graphical user session. > Combined with Pressured Exit aka idle-exit aka Sudden Termination, this > would allow your agents to run only when they are needed without worrying > about loading/removing their plists. > > > On Oct 18, 2016, at 08:32, Graham Miln <graham.miln(a)gmail.com> wrote: > > > > In macOS 10.12's ServiceManagement/ServiceManagement.h the following > functions are marked as deprecated since macOS 10.10: > > > > - SMJobSubmit > > - SMJobRemove > > - SMCopyAllJobDictionaries > > > > The header mentions replacements will be provided by libxpc. Are > replacements available or on the way? > > > > I have a system wide daemon that would like to programmatically launch a > process in a graphical user session. At the moment, our LaunchAgents > managed helper process is always running. Ideally, it would be good to > launch and exit the helper on-demand. > > > > I believe this can now be done with `launchctl bootstrap|bootout > gui/<uid>` on the command line. Should the daemon call out to the launchctl > tool? > > > > Kind regards, > > > > Graham > > _______________________________________________ > > launchd-dev mailing list > > launchd-dev(a)lists.macosforge.org > > https://lists.macosforge.org/mailman/listinfo/launchd-dev > > _______________________________________________ > launchd-dev mailing list > launchd-dev(a)lists.macosforge.org > https://lists.macosforge.org/mailman/listinfo/launchd-dev >

1 0

Replacements for deprecated SMJobSubmit and SMJobRemove
by Graham Miln 18 Oct '16

18 Oct '16

In macOS 10.12's ServiceManagement/ServiceManagement.h the following functions are marked as deprecated since macOS 10.10: - SMJobSubmit - SMJobRemove - SMCopyAllJobDictionaries The header mentions replacements will be provided by libxpc. Are replacements available or on the way? I have a system wide daemon that would like to programmatically launch a process in a graphical user session. At the moment, our LaunchAgents managed helper process is always running. Ideally, it would be good to launch and exit the helper on-demand. I believe this can now be done with `launchctl bootstrap|bootout gui/<uid>` on the command line. Should the daemon call out to the launchctl tool? Kind regards, Graham

2 1

Re: [launchd-dev] Mysterious problem with [NSURL getResourceValue:forKey:error:]
by James Bucanek 25 Aug '16

25 Aug '16

> Quinn "The Eskimo!" <mailto:eskimo1@apple.com> > August 23, 2016 at 5:27 AM > > Its sounds like you’re well on the way to a solution but I’m curious > about the above. In general the `UserName` property for a launchd > daemon is for role accounts. Putting a user account there is weird > because what user do specify? If there are two users simultaneously > logged in (via fast user switching, or truly simultaneously via screen > sharing), surely one of them has to miss out. Each daemon is named after its user: com.qrecall.switchboard.501, com.qrecall.switchboard.502, and so on. Each user connects only with its switchboard, and that switchboard only manages endpoints belonging to that user. Again, the *only* reason this is installed as a system daemon is that I have root processes (executed from another system daemon) that have to connect with it, and the only mach services namespace that I can make that work is the system's. James P.S. I would be nice if the UserName launchd property would also accept a number and interpret it as a UID: <key>UserName</key><integer>501</integer>.

2 2

Mysterious problem with [NSURL getResourceValue:forKey:error:]
by James Bucanek 23 Aug '16

23 Aug '16

Greetings, I've stumbled across a strange problem with [NSURL getResourceValue:forKey:error:] returning blank icons when executed within the context of a system daemon. I'm using this method to get the NSURLEffectiveIconKey (an NSImage) value for items. The process is launched by a privileged helper (installed a la EvenBetterAuthorizationSample). It's a SetUID executable, so its effective UID is the user (let's say 501) for which I want to get the icons (and other localized information). When the process is launched by the privielged helper, every NSImage returned by -getResourceValue:forKey:error: is a valid image, but the icons are all blank. When the process is launched from the shell (or via Xcode) as the regular user, the returned NSImages contain the correct icons. If the process is launched via sudo from the user's shell, the returned NSImages are also OK. I suspect that this is somehow a security or bootstrap context issue, but I'll defer to expert opinion on the cause or what can be done about it. James Bucanek

2 5

100% repeatable Kernel panic caused by launchd
by Camden Narzt 01 Jun '16

01 Jun '16

I was pointed at this mailing list to try and figure out what seems to be a pretty bad launchd bug. I recently added an opentsdb formula to homebrew, including a launchd plist. However when I load the plist with launchctl load ~/Library/LaunchAgents/homebrew.mxcl.opentsdb.plist my system hangs, then kernel panics, then on reboot it just hangs and doesn’t ever finish booting, requiring me to boot into single user mode, fsck the filesystem, mount the fs as read/write and then delete the plist. This is 100% reproducible on my MacBook Pro (Retina, 15-inch, Mid 2014) running OS X 10.11.5 (15F34). Steps to reproduce (warning, this will almost certainly cause a kernel panic): /usr/bin/ruby -e "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/master/install)" brew install opentsdb ln -s `brew --prefix hbase`/homebrew.mxcl.hbase.plist ~/Library/LaunchAgents/ ln -s `brew --prefix opentsdb`/homebrew.mxcl.opentsdb.plist ~/Library/LaunchAgents/ launchctl load ~/Library/LaunchAgents/homebrew.mxcl.hbase.plist sleep 15 && launchctl load ~/Library/LaunchAgents/homebrew.mxcl.opentsdb.plist Note that running hbase via launchd and starting opentsdb manually does not cause a kernel panic (or any other problems that I’ve noticed). Does anyone have any ideas about why this is happening?

7 9

Extending Bonjour information
by Mitchell Laurren-Ring 20 Apr '16

20 Apr '16

Within my launchd.plist I’m using the Bonjour key for some of my on-demand sockets. Is there a way to associate a TXT record with the name that gets automatically registered with the responder? An API I can call when launched but the info persists after my process has ended? /Mick

2 1

About the ExitTimeOut key in a daemon's configuration plist
by Axel Luttgens 19 Apr '16

19 Apr '16

Hello, From launchd.plist(5): ExitTimeOut <integer> The amount of time launchd waits between sending the SIGTERM signal and before sending a SIGKILL signal when the job is to be stopped. The default value is system-defined. The value zero is interpreted as infinity and should not be used, as it can stall system shutdown forever. Since I’m considering a service that may sometimes need somewhat more time than the default wait time (10 seconds, IIRC), I tried with: <key>ExitTimeOut</key> <integer>60</integer> in the service’s plist. No problem with a "launchctl bootout": the daemon receives a TERM signal and shuts down cleanly. But when rebooting the machine, it seems that the system doesn’t send a TERM signal: it just waits for sixty seconds, then KILLs the daemon. On the other hand, without the ExitTimeOut key in the plist, the TERM signal is sent to the daemon upon reboot (as expected). This is with Mac OS X 10.11.3. Am I misunderstanding the role of that ExitTimeOut key? TIA, Axel

2 4