[kami]

I wanted to report two new small bugs:

1 - When you have a scene rendering with three nodes and you release two of them (eg. at SL12). Then the combined SL of the job drops down and shows now exactly the same as the current SL of the one remaining node. If you make this job now stop at SL 15 for example, it merges all the mxi files of the three nodes and reaches a higher SL as shown in the monitor. (in my case it reached almost 17 in case of 15)

2 - When you have a job running with one node and it shows you that it still takes 1 hour to finish the job. Now you add one more node, the time drops down instantly and says that the job will finish in a few minutes. But this time the time counts up and gets more and more - It won't show a more or less correct value for a pretty long time.

[numerobis]

But this time the time counts up and gets more and more - It won't show a more or less correct value for a pretty long time.

It's the same as with resumed jobs. And with normal coop jobs it starts at high values counting down for a long time...

[kami]

It's the same as with resumed jobs. And with normal coop jobs it starts at high values counting down for a long time...

Yes, I noticed it with resumed jobs too.
The counting down with normal jobs doesn't bother me too much as usually you can kind of estimate pretty quickly how long it is really going to take after a few minutes. And it works the same as in the normal render engine - it finishes much quicker than the first time assumption.

[kami]

one more thing. my monitor loses pretty often the connection to the manager. sometimes I get this message:

There are error messages in the log starting at:[18/February/2014 09:53:58] TCP message from manager is not a XML message! You can find more info using the search tools in the log window.

sometimes it just doesn't update anymore without showing a message. not really a problem, as I can always restart the monitor, but it is a bit annoying.

Besides that, I wanted to thank you for improving the stability of the manager in v3. I didn't have any unexpected crashes (without an error message) anymore as before. And I've never managed to render so many jobs without having to restart the manager ... currently on job 169

[kami]

I made it until job 323, before I had to relocate the machine and reboot it for that
Never was that high before ...

Will the other stuff be fixed with the next beta?

Thx, Kami

[kami]

Sorry, I have to write it again:
It is really a huge problem, that the node (or manager) just closes if you click okay on an error message.
I now lost several hours of render time again for a project which has to be finished today, and had to restart several computers to reset the network. And this only because the was an older error message on the manager (because of some canceled project) and somebody clicked on that message, which closed the manager which was otherwise running fine. Of course restarting the manager would now crash the node on all computers that were rendering before ...
WHY?

[kami]

The next bug in the network render:
When you launch a job with SL14, but later you realize that you need it SL18, you can change that, with edit.
If you've done that and have for example 10 machines assigned to the job, it works fine. But none of the nodes renders higher than to SL14. So if you only have one machine assigned, the job will not come further than SL14 anyway.

Same goes for animation frames ... if you change the SL of the animation, after the first frames have started to calculate, these frames will never reach the new SL.

[Mihai]

If you've done that and have for example 10 machines assigned to the job, it works fine. But none of the nodes renders higher than to SL14. So if you only have one machine assigned, the job will not come further than SL14 anyway.

This is a bit confusing the way you put it.....so basically you are saying the Edit SL doesn't work, if you set it to a higher SL than specified in the original MXS?

[kami]

basically yes. it does work if you have enough machines assigned to the job.

sorry, if I explained it a bit complicated. but try this:
- start a rendering with SL10 on just one machine, edit the SL to 18. the rendering will stop at 10.

and as an extra:
- start a rendering with SL10 on two machines, edit the SL to 18. the rendering on both nodes will stop at 10, giving you a combined SL of around 12
etc ...

[kami]

Another thing:
I resumed an mxi (via network) to let it calculate from SL12 to SL18. He failed to merge and when I tried to merge the files manually he said, that he couldn't merge files with the same CPU id.
How is that possible? Just bad luck? It's not the first time that has happened to me, although it does not happen extremely often.