September 2009 - launchd-dev - lists.macosforge.org

Privilege separation and crash resistance
by Thomas Clement 07 Oct '09

07 Oct '09

Hello, I think the following architecture is recommended when a daemon needs to talk to an agent: - daemon listens on privileged port - agent is set to run at load time and checks in with the daemon to vends its mach port - daemon uses the vended mach port to launch the agent on demand However if the daemon crashes, it loses the vended mach port and is unable to launch the agent. What is the recommended solution to this situation? Regards, Thomas

3 5

Daemon logging and Snow Leopard
by Thomas Clement 23 Sep '09

23 Sep '09

It seems there are some inconsistencies about daemon logging on Snow Leopard: - asl(3) logs to the log database and system.log - NSLog logs only to system.log - printf logs nowhere The documentation states that NSLog uses asl(3), however these logs do not appear in the log database. Also the "Daemons and Agents" tech note states: > Logging With printf > In Mac OS X 10.5 and later, launchd will capture any output to > stdout and stderr and redirect it to ASL. However I fail to see any printf reaching the log database or even system.log. Why is that? Thanks, Thomas

1 0

Re: [launchd-dev] launchd-dev Digest, Vol 27, Issue 8
by Nehemiah Dacres 22 Sep '09

22 Sep '09

alright, at your first post I though you were reloading program code while a progam is running. Instead your changing launch properties of the program via the program being launched while its running. the best way to do this via the program is to edit the file using the cocoa framework instead of a script (documentation here<http://developer.apple.com/mac/library/DOCUMENTATION/Cocoa/Conceptual/Prope…>) and using the system() call from stdlib "system("launchd load com.yourdomain.yourprogram); " I still have no idea why you think the program should be changed by the server instead of a client but whatever. such a repetitive, brainless task should be left as stateless as possible to allow easy debugging. when you change things programmaticly, if that change isn't absolutely predictable, things get unstable quickly.

1 0

Re: [launchd-dev] launchd-dev Digest, Vol 27, Issue 5
by Sidney San Martín 18 Sep '09

18 Sep '09

How would a WatchPath help the situation? launchd already starts me when something tries to connect to the socket. launchd only reads your plist at startup/login, so changing the StartInterval on disk will do nothing (unless you know something I don't). Currently, if modifying the job in memory isn't possible, I think the best options are to either run all the time and manage timings myself, or to modify the StartInterval on disk and fire up a helper tool to unload/reload my plist (which really should not happen too often). If clients are connected, they'll be booted out for a few seconds and reconnect themselves. No big loss. On Fri, Sep 18, 2009 at 2:48 PM, Nehemiah Dacres <vivacarlie(a)gmail.com> wrote: > so you want the ability to connect to it ALL the time but dont want it > running All the time. note that a waiting process doesn't take that much > proessor time if its not doing anything but listening to a socket but if you > must you can have the client program modify a file on a path and set the > server's launchd.plist for "WatchPaths" to start the daemon when said path > is altered. > I guess you can change startInterval or IdleTime from the program, the test > (which I may not perform later) is if the start interval will be changed if > the file is not reloaded unto launchd after altering. I would say > > On Fri, Sep 18, 2009 at 1:12 PM, Sidney San Martín <s(a)sidneysm.com> wrote: >> >> It won't. The daemon manages its timing while running, and would work >> (with no modifications) with KeepAlive and no OnDemand keys at all. >> I'd just rather NOT have the daemon idling away all day when launchd >> has a perfectly good mechanism for launching it periodically. >> >> On Fri, Sep 18, 2009 at 1:12 PM, Nehemiah Dacres <vivacarlie(a)gmail.com> >> wrote: >> > >> >> I see no reason not to let the daemon time out and exit until >> >> a client tries to connect or until its next checkin. >> >> >> > >> > neither do I, but who says the program will start up if its already >> > running? >> > >> > >> > -- >> > >> > "lalalalala! it's not broken because I can use it" >> > >> > >> > http://linux.slashdot.org/comments.pl?sid=194281&threshold=1&commentsort=0&… >> > > > > > -- > > "lalalalala! it's not broken because I can use it" > > http://linux.slashdot.org/comments.pl?sid=194281&threshold=1&commentsort=0&… >

1 0

Re: [launchd-dev] launchd-dev Digest, Vol 27, Issue 5
by Sidney San Martín 18 Sep '09

18 Sep '09

It won't. The daemon manages its timing while running, and would work (with no modifications) with KeepAlive and no OnDemand keys at all. I'd just rather NOT have the daemon idling away all day when launchd has a perfectly good mechanism for launching it periodically. On Fri, Sep 18, 2009 at 1:12 PM, Nehemiah Dacres <vivacarlie(a)gmail.com> wrote: > >> I see no reason not to let the daemon time out and exit until >> a client tries to connect or until its next checkin. >> > > neither do I, but who says the program will start up if its already running? > > > -- > > "lalalalala! it's not broken because I can use it" > > http://linux.slashdot.org/comments.pl?sid=194281&threshold=1&commentsort=0&… >

1 0

Re: [launchd-dev] launchd-dev Digest, Vol 27, Issue 5
by Nehemiah Dacres 18 Sep '09

18 Sep '09

That's very dangerous. > > How can I modify properties of a job (StartInterval, in this case) without un/reloading it from launchd? when I read "how do i modify x without reloading it" I see a design issue. Startinverval " When an agent has the StartInterval key in its .plist file, so that it is run every *N* seconds" from macgeekery . which is a summary of the launchd manual *StartInterval* *<integer>* This optional key causes the job to be started every N seconds. If the system is asleep, the job will be started the next time the computer wakes up. If multiple intervals transpire before the computer is woken, those events will be coalesced into one event upon wake from sleep. Ideally, I want a daemon to be > able to modify its own StartInterval while running, so that it applies > the next time the daemon idles out, and without disrupting any of its > launchd-provided sockets. > > Idleingout, I think referes to OnDemand, because you want to control when the job comes back after being stopped from lack of use? This is probibly not necessary. again from the man page *KeepAlive* *<boolean* *or* *dictionary* *of* *stuff>* This optional key is used to control whether your job is to be kept continuously running or to let demand and conditions control the invocation. The default is false and therefore only demand will start the job. The value may be set to true to unconditionally keep the job alive. Alternatively, a dictio-nary dictionary nary of conditions may be specified to selectively control whether *launchd* keeps a job alive or not. If multiple keys are provided, launchd ORs them, thus providing maximum flexibility to the job to refine the logic and stall if necessary. If *launchd* finds no reason to restart the job, it falls back on demand based invocation. Jobs that exit quickly and frequently when configured to be kept alive will be throttled to converve system resources. *SuccessfulExit* *<boolean>* If true, the job will be restarted as long as the program exits and with an exit status of zero. If false, the job will be restarted in the inverse condition. This key implies that "RunAtLoad" is set to true, since the job needs to run at least once before we can get an exit status. *NetworkState* *<boolean>* If true, the job will be kept alive as long as the network is up, where up is defined as at least one non-loopback interface being up and having IPv4 or IPv6 addresses assigned to them. If false, the job will be kept alive in the inverse condition. *PathState* *<dictionary* *of* *booleans>* Each key in this dictionary is a file-system path. If the value of the key is true, then the job will be kept alive as long as the path exists. If false, the job will be kept alive in the inverse condition. The intent of this feature is that two or more jobs may create semaphores in the file-system namespace. *OtherJobEnabled* *<dictionary* *of* *booleans>* Each key in this dictionary is the label of another job. If the value of the key is true, then this job is kept alive as long as that other job is enabled. Otherwise, if the value is false, then this job is kept alive as long as the other job is disabled. This feature should not be considered a substitute for the use of IPC. Tell us why you want to do that and I could probibly give you a better design decision. -- "lalalalala! it's not broken because I can use it" http://linux.slashdot.org/comments.pl?sid=194281&threshold=1&commentsort=0&…

2 1

Re: [launchd-dev] Recommended tool daemon/agent install folder ... Is this list still active?
by Karl Moskowski 17 Sep '09

17 Sep '09

On 2009-09-16, at 3:25 PM, Brent Gulanowski wrote: > Now I'm curious if anyone here remembers the standard path for tools > on Mac > OS X which are launched by launchd -- there's a folder in Library > which > isn't created by default but is considered proper, but I can't > recall nor > find any mention of it. Brent, Apple's BetterAuthorizationSample uses /Library/ PrivilegedHelperTools/, and that location seems to have become a de facto standard. <http://developer.apple.com/samplecode/BetterAuthorizationSample/> Caveats: - ensure any installers keep the correct permissions and ownership on that folder; - ensure your tool is uniquely named, probably using reverse DNS notation (e.g., com.mycompany.foobard). (There are probably more.) ---- Karl Moskowski <kolpanic(a)voodooergonomics.com> Voodoo Ergonomics Inc. <http://voodooergonomics.com/>

1 0

launchd daemon launch delayed when (Snow) Leopard Server Firewall is on (port is open)
by Brent Gulanowski 17 Sep '09

17 Sep '09

We've found a problem with Daylite Server that if we turn on the Leopard Server Firewall and open the Daylite Server ports for incoming connections, launchd doesn't seem to get notified, or at best there is a long delay, which exceeds our timeout. Longer than two minutes. I have confirmed that launchd does eventually wake up by connecting with telnet. Turning the firewall off results in normal behaviour. I'm not sure where to start to identify the source of the problem. I could not find any other reports of this exact behaviour. If anyone has any thoughts or can point me in the right direction, that would be appreciated. Here is the property list as it is found in /Library/LaunchDaemons: <?xml version="1.0" encoding="UTF-8"?> <!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd"> <plist version="1.0"> <dict> <key>GroupName</key> <string>_dayliteserver</string> <key>Label</key> <string>com.marketcircle.daylite.dlserver</string> <key>ProgramArguments</key> <array> <string>/Library/Application Support/Daylite Server/Daylite Server.bundle/Contents/MacOS/DLServer</string> </array> <key>Sockets</key> <dict> <key>Listeners</key> <dict> <key>Bonjour</key> <string>dlserver</string> <key>SockServiceName</key> <string>6113</string> </dict> </dict> <key>StandardErrorPath</key> <string>/Library/Application Support/Daylite Server/Logs/Daylite Server.log</string> <key>UserName</key> <string>_dayliteserver</string> <key>inetdCompatibility</key> <dict> <key>Wait</key> <false/> </dict> </dict> </plist> -- Brent Gulanowski

3 2

Recommended tool daemon/agent install folder ... Is this list still active?
by Brent Gulanowski 17 Sep '09

17 Sep '09

I posted a question a week or so ago, but either I used the wrong address or no one noticed. Haven't seen any messages on this list since I subscribed. My last question was about a lauchd job not starting immediately when the firewall on OS X Server was on. Now I'm curious if anyone here remembers the standard path for tools on Mac OS X which are launched by launchd -- there's a folder in Library which isn't created by default but is considered proper, but I can't recall nor find any mention of it. Thanks, -- Brent Gulanowski

3 2

Possible to modify a job in-memory?
by Sidney San Martín 15 Sep '09

15 Sep '09

How can I modify properties of a job (StartInterval, in this case) without un/reloading it from launchd? Ideally, I want a daemon to be able to modify its own StartInterval while running, so that it applies the next time the daemon idles out, and without disrupting any of its launchd-provided sockets.

1 0