Hello !
It’s been some updates, i’ve lost my pulse connection from slave.
Everything used to work fine, that’s why i didn’t worried about it.
In monitor : slave are mark ‘Connected to pulse : No’, but each node can ping my Pulse machine.
Every pulse features seems to work : stats, throttling…
But since i’ve 150+ nodes up & running, things seems slower, more nodes are going stalled.
Is it a monitor problem ? How can i check if a slave is really connected to pulse ?
Thanks
-p-
If this is Deadline 8, the Slaves no longer need to connect to Pulse as they now use the Database for throttling. We forgot to remove the column, but the Slaves will never say “Yes” to being connected to Pulse.
As far as the slowness and stalling, that is definitely a bug. Could you send along a few Slave reports? They should not be stalling out.
yes it’s 8.0.6.5
Thanks for confirming that this column is useless, i was suspecting that.
Global slowness & stalled machines certainly comes from the fact that we entered in full production with heavy 3D jobs 
-p-
Hah, have to say I was wondering about this since our upgrade to DL8 about month ago.
It’ll be gone in 8.1 I believe. It’s definitely thrown a few people for a loop since the throttling change is maybe a one line mention in our release notes.
It was only ever used for throttling after 6.0 (when we removed the queuing/caching smarts from Pulse), so nothing else should have been affected.
So does Pulse has any interest anymore ?
It’s great for power management and cleaning things up so the Slaves don’t have to.
For example, you can have Pulse stop machines if the farm is idle or if your sever room is too hot (assuming a supported temperature sensor), then turn the machines on again once new jobs come in.
Considering the web service was moved to a standalone app and throttling is handled in the DB, Pulse really just handles those things.
For me, the most important thing that end users ask a million times about is for a machine to regularly carry out the “Pending Job Scan”. If your entire farm is busy (which it should be, if you are trying to make art/money 
), then your Slaves are busy rendering, so they are unable to be idle, which is when they carry out the pending job scan.
Running Pulse, it will carry out this Pending Job Scan for you at regular intervals and then end users are super happy, as pending jobs get released in a timely fashion for them.
Pulse is your friend. Always run Pulse, not for performance reasons, but because it’s super useful.
Thanks guyz !
Indeed, we are rendering nearly all the time, so the nodes are never idle, and we don’t have WoL or T°C sensor up and running.
We have 10-20 artists always managing things with monitor, so i think the pulse get all this a bit more smooth.
Last question : my sysAdmin want to replace my actual pulse running on Windows server by a Linux one (all our network is windows). I suppose that there’s no conflits. Is there a interest of having 2 pulse ? more is always better ? 
Thanks !
-p-
Sure. Pulse runs just fine and identical on the 3 x platforms we support - Win, Linux, OS X. Running multiple Pulses is a good idea for redundancy as well: docs.thinkboxsoftware.com/produc … -ref-label
One is set to be the Primary and the others act as Secondary, waiting to take over if the primary goes down.