Asset Crawler Server

Deadline
1

Using VMX Balancer Options found in Configure Cloud Providers dialog it is possible to enable or disable Asset Checking, specify Crawler Hostname, Port and OS.

642×603

Then there is AssetCrawler_Server.py script in Drive:\DeadlineRepository8\vmx directory which is as its file name suggests is a server. It is probably intended to be started on a cloud render slave as its job’s pre-render process.

AssetCrawler_Server’s HandleConnection expects data argument:

data = conn.recv( BUFFER_SIZE )

What data and what data type/format the data is?
If AssetCrawler_Server is indeed started on a slave machine as a pre-render process than what process is responsible for collecting the Assets data (Asset file paths?) and calling the AssetCrawler Server process started on slave?

2

The Asset Crawler could use a bit more love, but as a server it is only intended to run on one machine per region. The idea is that it walks over the file servers and keeps track of what files exist within the region. It’s faster to do that check within the particular network, and then the Balancer can query if the asset(s) it needs are available before spawning an instance.

I believe this is driven by a job’s asset dependency list, and the Slaves themselves should be able to check for those assets without the Crawler.

We don’t have any support scripts for it such as an init script or a plist, and I haven’t worked with it much. Are you using the balancer heavily? I can likely do a deep-dive revisit if you need it.