November 2007 - launchd-dev - lists.macosforge.org

Run job when network configuration changes
by Louis Granboulan 13 Dec '07

13 Dec '07

Dear all, I would like to support Jeremy Reichman's request, because I apparently have similar needs. I have used the Kicker.xml technique to make automatic backups of my laptop. However, this laptop connects to various networks, and should backup only when connected to the ones where enough bandwidth is available. Moreover, one of such networks uses a 802.1x authentication, and I can backup only when authenticated. In that case, I go from a wired- non-authenticated connection to a wired-authenticated connection, and I need to launch the backup script after authentication. Thanks, Louis

8 16

launchctl unload not working
by James Bucanek 30 Nov '07

30 Nov '07

Greetings, I'm having a number of issues with launchd, particularly under Leopard. (Most, if not all, of this works just fine under Tiger.) My current, and most pressing issue, is trying to replace a daemon when the user upgrades their software. Here's what I've got running - A daemon is installed in /Library/LaunchDaemons/QRecallScheduler501.plist. I install the file by writing the plist then executing 'launchctl load /path-to-plist' as root. The .plist file has the UserName property set to an individual user. This has been working (mostly) until I need to upgrade the user's software. I've tried various techniques, but the one I'm using now which I think should work doesn't. Setup (if this makes any difference): The path to the daemon is a symbolic link to the actual binary in my application bundle. When the application is upgraded (via Sparkle), the application bundle is replaced with the new application bundle containing an updated daemon. Step #1: Run 'launchctl unload -w /path-to-plist' as root. I wait for launchctl to return and then wait for the daemon to terminate. Step #2: Delete the .plist at /Library/LaunchDaemons/QRecallScheduler501.plist. Step #3: Write a new .plist at /Library/LaunchDaemons/QRecallScheduler501.plist Step #4: Run 'launchctl load /path-to-plist' as root. Wait for launchctl to finish. Here's what happens: In step #1, launchctl returns with a status of 0. The daemon receives a SIGTERM and shuts down (2-3 seconds). IMMEDIATELY, launchd starts the new deamon. I can only assume because the location of the updated binary is the same as the old one that was just terminated. Steps #2 and #3 delete and rewrite the .plist. Step #4 runs 'launchctl load' which returns with a status of 0. What I get in the console log is 11/28/07 4:09:44 PM [0x0-0x1e01e].com.qrecall.client[145] launchctl: Error unloading: com.qrecall.scheduler 11/28/07 4:09:56 PM com.apple.launchd[62] (com.qrecall.scheduler.501) Ignored this key: UserName Contrary to the second error, the daemon is run with a UID of 501 (which is correct -- the daemon will terminate if it's started as root). The new daemon is then repeatedly restarted. (The daemon checks for a duplicate running instance of itself and terminates immediately if it's already running.) 11/28/07 4:09:57 PM com.apple.launchd[62] (com.qrecall.scheduler.501) Throttling respawn: Will start in 10 seconds 11/28/07 4:10:07 PM com.apple.launchd[62] (com.qrecall.scheduler.501) Throttling respawn: Will start in 10 seconds 11/28/07 4:10:17 PM com.apple.launchd[62] (com.qrecall.scheduler.501) Throttling respawn: Will start in 10 seconds 11/28/07 4:10:27 PM com.apple.launchd[62] (com.qrecall.scheduler.501) Throttling respawn: Will start in 10 seconds The really bad part is the instance of the daemon that is running does not seem to be running in the correct environment/namespace. When it launches a sub-process it attempts to connect with it using Mach ports; The connection fails. Restarting the OS doesn't fix the problem. I think the problem is that (in Leopard) there seems to be two instances of launchd running: One as root and one as user 501. But I'm getting very confused as to which one I should be trying to deal with. I changed the code to call launchctl while running as user 501. That sometimes works and sometimes doesn't. If I can get the daemon to stop and issue 'launchctl load /...' as user 501, the daemon starts working: specifically, it can communication with the sub-processes that it starts. Once started using 'launchctl' as user 501 it can be stopped again using launchctl as 501. But after a restart, it's running in the wrong namespace again and I have to perform a 'sudo launchctl unload' to get it to stop. So after an entire day, I'm really confused as to what launchctl commands I should be issuing to replace a running deamon, in what order, and using what UID. James -- James Bucanek

5 14

User Deamon
by James Bucanek 27 Nov '07

27 Nov '07

Welcome, I was invited to bring this thread here by Quinn. My apologies if you've seen bits and pieced of this thread before; it's traveled around the Cocoa, Carbon, and MacNetworkProg lists before landing here. I'm developing a backup and document archiving solution (<http://www.qrecall.com/> -- not an ad, just for reference). The program provides "actions" that can be scheduled to run at regular intervals or in response to certain events. These actions should be performed whether the user is logged in or out. In fact, actions can be scheduled to run only when the user is logged out or because the user logged out. There is a scheduler process which is responsible for managing all of this. The scheduler works on behalf of a single user. (Other users may have their own schedule/scheduler.) For security, the scheduler runs under the UID of that user. So, what I have is a process that - Needs to start up when the system boots and runs until the system shuts down, independent of whether the user is logged in or not. - Runs as a single user (not as root) without any special privileges. - Can be installed and removed without requiring administrative authorization. In effect, I need a "User Daemon" -- a class of process that launchd does not provide. I was simulating this (pretty well) using crontab. The crontab will allow non-admin users to install a cronjob that runs periodically and independently of their login session. Unfortunately, recent versions of Tiger and Leopard have become downright hostile towards user cronjobs, sending them SIGKILL signals whenever the user is logged out (without even the courtesy of a SIGTERM first!). I'm now simulating this by installing a system-wide daemon for each user (/Library/LaunchDaemons/QRecallScheduler501.plist, .../QRecallScheduler502.plist, etc.). Each daemon is configured to run as that user (via the UserName property). But this requires administrative privileges to install/uninstall a non-privileged process. Beyond being annoying, it limits the usefulness of the application for non-admin users. And in the pit of my stomach, I feel that it presents a security hole. I'm basically just throwing this out to the launchd team for comments, which might result in filing a feature request. I also understand some of the potential problems in implementing user daemons. (For example, the logical place to place a user daemon configuration file would be in ~/Library/LaunchDaemons, but that obviously won't work with FileVault or users who mount their home folder over a network.) However, there are also potential solutions to some of these and there's also the principle that one shouldn't necessarily eliminate a feature just because 0.1% of the users can't use it. ;) Any thoughts? -- James Bucanek

2 2

Re: [launchd-dev] Getting user parameters as an agent
by Yaniv Aknin 19 Nov '07

19 Nov '07

My apologies, I'm unable to reproduce the behavior I saw earlier with 'user-specific' DAApprovalSessionRef. I haven't the faintest how it happened, maybe a bug in some other part of my code and I jumped to conclusions. Anyhow, if I hit an obstacle, I'd re-raise the issue with the list. Still, I'd be interested in knowing how can I provide username based variables in launchd's plists, if at all possible. Thanks and apologies again, - Yaniv On 11/19/07, Quinn <eskimo1(a)apple.com> wrote: > > At 23:14 +0200 17/11/07, Yaniv Aknin wrote: > >I'm writing a Tiger agent in Python and C which must run as root (it > >enforces policies on the user). The agent should, however, know > >which user it is enforcing (this is due to DiskArbitration's > >inability to create a 'global' Approval session affecting all users). > > It seems like this is where you're getting into trouble. DiskArb is > a daemon-safe framework. I believe it should be possible for a > daemon to implement a DAApprovalSessionRef that affects all mounts on > the system. Can you explain how you tried to do that and what went > wrong? > > S+E > -- > Quinn "The Eskimo!" < http://www.apple.com/developer/> > Apple Developer Relations, Developer Technical Support, Core OS/Hardware > _______________________________________________ > launchd-dev mailing list > launchd-dev(a)lists.macosforge.org > http://lists.macosforge.org/mailman/listinfo/launchd-dev >

2 1

Getting user parameters as an agent
by Yaniv Aknin 19 Nov '07

19 Nov '07

Hi, I'm writing a Tiger agent in Python and C which must run as root (it enforces policies on the user). The agent should, however, know which user it is enforcing (this is due to DiskArbitration's inability to create a 'global' Approval session affecting all users). I've seen references on the list to the fact that an Agent in Tiger absolutely must run as the user, no workarounds. This is pretty bad for me, and would force me to do something crazy like running a daemon as root and installing a tiny agent notifying the daemon a user came into existence (I'd rather not link with Cocoa notifications). Is there any way around that? Something saner than what I was suggesting? Also, there's the issue of having launchd tell my agent which user it's running for, either on the command line, as an environment variable, or otherwise. And if I'm not asking for too much (I know I am...) - also redirect stdout to a user-specific locations (/var/log/my-app/log.$USERNAME). How should I go about doing that? I couldn't find something relevant in launchd.plist's manpage nor on the list's archive. Many thanks, - Yaniv

2 1

Attaching to launchd via gdb?
by Matt DeFoor 16 Nov '07

16 Nov '07

Is there a means to attach to and debug launchd? Or to attach to a per user launchd process? I get the following when logged in as root and trying to attach to launchd: (gdb) attach launchd.1 Attaching to program: `/sbin/launchd', process 1. Unable to attach to process-id 1: Operation not permitted (1). This request requires that the target process be neither setuid nor setgid and have the same real userid as the debugger, or that the debugger be running with administrator privileges. I'm trying to determine why the call to getpwuid() in job_postfork_become_user() is failing. Here's a snippet from system.log: Nov 15 14:39:35 dc3 com.apple.launchd[1] (com.apple.launchd.peruser.166128603[1036]): getpwuid("166128603") failed Nov 15 14:39:35 dc3 com.apple.launchd[1] (com.apple.launchd.peruser.166128603[1036]): PID 1002 "authorizationhos" has no account to back it! Real/effective/saved UIDs : 0/166128603/0 Nov 15 14:39:35 dc3 com.apple.launchd[1] (com.apple.launchd.peruser.166128603[1036]): PID 112 "authorizationhos" has no account to back it! Real/effective/saved UIDs: 0/166128603/0 Nov 15 14:39:35 dc3 com.apple.launchd[1] (com.apple.launchd.peruser.166128603[1036]): Exited with exit code: 1 Nov 15 14:39:35 dc3 com.apple.launchd[1] (com.apple.launchd.peruser.166128603): Throttling respawn: Will start in 10 seconds Cheers, Matt

2 2

Environment variables
by Hamish Allan 15 Nov '07

15 Nov '07

Hi, The man page for launchd.plist says: SecureSocketWithKey <string> This optional key is a variant of SockPathName. Instead of binding to a known path, a securely generated socket is created and the path is assigned to the environment variable that is inherited by all jobs spawned by launchd. However, the SSH_AUTH_SOCK environment variable generated by Leopard's /System/Library/LaunchAgents/org.openbsd.ssh-agent.plist doesn't seem to be available to my ~/Library/LaunchAgents. E.g. the following: <?xml version="1.0" encoding="UTF-8"?> <!DOCTYPE plist PUBLIC "-//Apple Computer//DTD PLIST 1.0//EN" "http:// www.apple.com/DTDs/PropertyList-1.0.dtd"> <plist version="1.0"> <dict> <key>Label</key> <string>test.7B21D717-E7AE-4DB0-905F-7D40F8A6ABE2</string> <key>Program</key> <string>/usr/bin/env</string> <key>inetdCompatibility</key> <dict> <key>Wait</key> <false/> </dict> <key>Sockets</key> <dict> <key>Listeners</key> <dict> <key>SockNodeName</key> <string>127.0.0.1</string> <key>SockFamily</key> <string>IPv4</string> <key>SockServiceName</key> <string>34567</string> </dict> </dict> </dict> </plist> when accessed via 'nc localhost 34567' generates a list of several environment variables, none of which is SSH_AUTH_SOCK. Is this the expected behaviour? If so, how can I make a LaunchAgent for ssh that can access my ssh-agent across reboots? Thanks, Hamish

2 4

Dynamic port listening
by Hamish Allan 13 Nov '07

13 Nov '07

Hi, Is it possible to make a LaunchAgent listen on the first free port rather than a specific one? I'm trying to forward an HTTP connection on demand (over SSH), but rather than choose a specific local port I'd rather just choose a free one and then advertise it using Bonjour. So I also need to know what the chosen port was and pass it to dns-sd. A related question: is it possible (without writing to /etc/services) to make LaunchAgent listen on a port decided dynamically? i.e. would it be possible to use NAT-PMP to forward an external port to the first available local port, and then have a LaunchAgent listen on that port? Thanks, Hamish

2 4

bitlbee - xinetd to launchd
by Johannes Kneer 10 Nov '07

10 Nov '07

Hi My problem boils down to porting this xinetd file to launchd: # default: off # description: Bitlbee is an IRC gateway to other networks. service ircd { type = UNLISTED socket_type = stream protocol = tcp wait = no user = bitlbee group = nogroup server = /sw/sbin/bitlbee port = 6667 disable = no log_on_failure += USERID } So this is what I have come up with: <?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>KeepAlive</key> <true/> <key>Label</key> <string>org.bitlbee.bitlbee</string> <key>OnDemand</key> <true/> <key>ProgramArguments</key> <array> <string>/sw/sbin/bitlbee</string> </array> <key>ServiceDescription</key> <string>bitlbee irc-im proxy</string> <key>Sockets</key> <dict> <key>Listener</key> <dict> <key>SockFamily</key> <string>IPv4</string> <key>SockProtocol</key> <string>TCP</string> <key>SockServiceName</key> <string>6667</string> <key>SockType</key> <string>stream</string> </dict> </dict> <key>inetdCompatibility</key> <dict> <key>Wait</key> <false/> </dict> <key>UserName</key> <string>bitlbee</string> </dict> </plist> And well it does not work. When I try to load (launchctl load thisfile.plist) I always get a "bind(): Operation not permitted" error and I can't connect to the server. And how would I restrict the service to localhost only? I guess I'm missing something. I'd just like to have bitlbee working in leopard. Would apreciate some help. Johannes

5 8

TN2083 "Daemons and Agents" update
by Quinn 06 Nov '07

06 Nov '07

Greetings All Just FYI, I've done an extensive overhaul of DTS Technote 2083 "Daemons and Agents" to account for many of the changes in process management in Mac OS X 10.5. That update is now on the Apple developer web site. <http://developer.apple.com/technotes/tn2005/tn2083.html> Share and Enjoy -- Quinn "The Eskimo!" <http://www.apple.com/developer/> Apple Developer Relations, Developer Technical Support, Core OS/Hardware

1 0