AWS Thinkbox Discussion Forums

Automatic upgrade issue

I’m having an issue with automatic upgrades from beta 12 to beta 13. From my testing, more often than not the upgrade will run, but then launcher won’t re-open automatically. If I watch task manager while this is happening, deadlinelaunchernew.exe starts, and then deadlinelauncher.exe closes as expected. But then when the update is done, deadlinelauncher.exe starts for less than a second, after which both processes close. If I look at the deadline logs, the latest updated log was from deadlinelaunchernew, which looks successful as far as I can tell:

2013-03-06 14:52:36: BEGIN - WS-17\slave
2013-03-06 14:52:36: Deadline Launcher 6.0 [v6.0.0.50272 R]
2013-03-06 14:52:36: deadlinelaunchernew is running
2013-03-06 14:52:46: Killing existing deadlinelauncher because it failed to exit after 10 seconds
2013-03-06 14:52:46: deadlinelauncher has exited, proceeding with self upgrade
2013-03-06 14:52:51: Reading in bootstrap ini settings…
2013-03-06 14:52:51: Bootstrap file: C:\Users\slave\AppData\Local\Thinkbox\Deadline6\temp\LauncherBootstrap\bootstrap.ini
2013-03-06 14:52:51: Copying \deadline\deadline-test\bin\Windows\bin.zip to C:\Users\slave\AppData\Local\Thinkbox\Deadline6\temp\LauncherBootstrap\bin.zip
2013-03-06 14:52:51: Copying \deadline\deadline-test\bin\Windows\Version to C:\Users\slave\AppData\Local\Thinkbox\Deadline6\temp\LauncherBootstrap\Version
2013-03-06 14:52:51: Decompressing C:\Users\slave\AppData\Local\Thinkbox\Deadline6\temp\LauncherBootstrap\bin.zip to C:\Program Files\Thinkbox\Deadline6\bin
2013-03-06 14:52:52: Copying C:\Users\slave\AppData\Local\Thinkbox\Deadline6\temp\LauncherBootstrap\Version to C:\Program Files\Thinkbox\Deadline6\bin\Version
2013-03-06 14:52:52: Launch Slave: True
2013-03-06 14:52:52: Launch Pulse: False
2013-03-06 14:52:52: Launch Monitor: False
2013-03-06 14:52:52: Launching launcher
2013-03-06 14:52:52: Self upgrade completed
2013-03-06 14:52:52: Shutting down

But then the new deadlinelauncher process seems to die right after it’s launched. If I start the launcher manually after that everything works (and the upgrade was successful). I can go back and forth between our production repository (beta 12) and our test repository (beta 13), and it happens every time on most (but not quite all) machines.

Do you think this is a bug of some sort, or would it be a problem on our end?

Thanks.

Darren

So when we were testing this locally, Edwin was having this exact same issue on his laptop. In spite of some pretty serious debugging efforts, we couldn’t figure out why the second launcher was just dying. Since his was the only machine running into the issue (worked fine on all the other ones we tested it on), we hoped it was some one-off weirdness – it looks like that’s not the case though, so we’ll have another go at figuring this out.

Great, thanks! If you need any more info about the machines that this is happening on, just let me know. All the machines in question (those that worked and those that didn’t) are running Windows 7 Professional 64-bit, and the user has local admin privileges.

FYI, I just went ahead with upgrading our main repository to beta 13, and this happened on well over half of the slaves and workstations. It’s not a big deal with the current size of our farm, but it did mean that I had to VNC in and manually start the slave on most of them.

So I’m pretty sure we figured this out – it seems to be an issue with the first launcher getting killed by dealdinelaunchernew.exe, and not cleaning up the Mutex we use to determine if another copy is running. The confusing part is that this doesn’t seem to cause issues on all machines, I still had to rely on Edwin to test stuff out for me.

Either way, this should be fixed when the time to upgrade to beta 15 comes around

Cheers,

  • Jon
Privacy | Site terms | Cookie preferences