What's going on with NR 7.1.51?

Started by monson67, October 13, 2017, 01:29:58 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

monson67

I downloaded the update to KS Network Render, and nothing works anymore. Slaves/Workers won't connect to the master. Also, the GUI looks identical to NR6. Can I get links to downgrade to 7.1.36?

Rex

Unfortunately there were some serious issues with the NR7 data store and authentication that have led us to revert back to the NR6 solution.

When updating from 7.1.36 to 7.1.51 you will have to run the update installer on every slave. Also please make sure the port values are correct.

If you continue to have issues please contact support@luxion.com

KeyShot

KS NR 7.1 was released to address some issues with our enterprise customers, that experienced some instability and scaling issues with the new network rendering on a large number of machines. After careful consideration, we decided to continue with the KS 6 NR architecture. KS NR 7.1 will work as a fresh install. It is not intended as an upgrade of KS 7.0 NR. We will release KS NR 7.2 which can be used to upgrade KS 7.0 NR. We will also add key features and optimizations found in KS NR 7.0 (that we removed in the 7.1 release). If you have specific features that you like please let us know, so we can prioritize getting those into NR 7.2

DMerz III

 :-\ hmm, we were/are big fans of the strides that were taken in 7.0

There were a lot of features in NR 7.0 that I really appreciated, as outlined in my raves and review post a month ago.
https://www.keyshot.com/forum/index.php?topic=19596.msg81898#msg81898

If you guys are revamping the release, things we loved;

+ Worker Pools and the ability to assign a job to a pool upon transfer. (this is a huge deal for us, please retain)

+ Job details, I loved the increased transparency of information about a job, such as the dimensions and the ability to clone a job and change a few parameters such as sample rate/image size.

+ Tasks, I liked being able to see how tasks were progressing, though I wish the 'timer' for each job would not start until the first region has begun rendering instead of starting when the job is transferred (this would help me gauge actual rendering time vs idle time).

+ Preview seemed to work a hell of a lot better in 7 than 6.

Things we would love to see;
= Ability to change the default worker pool (right now if you turn off default in preferences, you will get a prompt each time a new job is sent, in which case you need to choose a pool for each job. I love this feature but wish I could 'batch' a certain # of jobs to a certain pool. Would be a bit of a UI challenge I'm sure.

= Ability to change the region dicing size. Right now it is automated, and as I mentioned in a wish list post, would love to make smaller regions in order to get more machines chewing on a tough area of my image, making tasks multiplied but divided and conquered.



monson67

I was able to get Network Render working once I got all the workers/slaves, the master, and the base software upgraded to 7.1.51. Having just one instance of software not up to date brought everything down.

Quote from: DMerz III on October 16, 2017, 08:56:18 AM
Ability to change the region dicing size. Right now it is automated, and as I mentioned in a wish list post, would love to make smaller regions in order to get more machines chewing on a tough area of my image, making tasks multiplied but divided and conquered.

I would echo everything DMerz said. I, too, love the new Tasks tab, breaking down which worker is rendering which region. But above all, his recommendation for changing region size, and therefore the number of regions, is something I would love to see.

Gordon

I am thrilled to have the NR6 style interface back, I was not a fan of v7 with the exception of those features that DMerz  mentioned and the inability to move the data store to a different drive is killing me in v7.  I also agree with his and monson's 'dicing size' want to make more regions.
However in the future, any time a dot revision update is NOT backwards compatible with older slaves and queues there needs to be a big flashing warning screen that pops up.  Did not expect I'd have to then spend the entire morning updating all of my slave machines and running around to every user in my company updating their NR version.