Hi,
I’m trying to virtualize a rendering node, but I’m facing a strange problem : in Deadline monitor the mac address is reported as FF:FF:FF:FF:FF:FF and there’s 2 lines in the slave list widget concerning that host (one tagged “vfarmer”, the other “vfarmer-localhost” in “Slave Name” column).
I use a bridged network to access my virtual OS (I guess that’s the problem). Is there anything I can do to make it work properly ? I’ve spent a lot of time configuring the VM image and I’m totally stuck at this point !
Best regards.
Cedric PAILLE
Head of R&D @ DwarfLabs
Hello,
The MAC address of a machine shouldn’t matter to Deadline, as we use hostname or IP address for connections. What issues are you running into because of this? I could see no issues on our end for our virtual machines.
Indeed, I wasn’t able to connect slave logs, but iptables was the problem on the vm side.
That said, I always have 2 instances of deadlineslave after having started the service, is this normal ?
Thank you.
Unfortunately, that’s a known bug on Linux and OS X if the Slave didn’t close properly. There is an option in /var/lib/Thinkbox/Deadline6/deadline.ini called “RestartStalledSlave”. Disable that to fix it. The line should look like so:
RestartStalledSlave=False
Here are the details:
The feature is built into the Launcher to check and see if the Slave is in a stalled state, and restarts it if it is. Now, it doesn’t check if it had just started the Slave for normal startup reasons so it essentially asks two to begin. There’s a second bug in 6.2 on Mac and Linux that this uncovered where if Slaves start close enough to each other, they don’t detect that one is running and bail out properly.
Both of these are fixed in the code base, and we want to get a service pack out to address these. We’ll be releasing to the beta boards soon, but if you want to test drive it yourself, e-mail me at edwinamsler@thinkboxsoftware.com and I’ll get you hooked up when it becomes available in the next few weeks.
Thank you for the information, I’ve found some files with a “.ini” extensions in the “slave” configuration directory, all is working fine after having deleted them, problem solved !
Best regards
Ah, right. Sorry I forgot to ask if the names were the same or not.