ok thanks! i did follow instructions and this is result see attachments...abstrax wrote:You would follow these instructions3dworks wrote: thank you for getting back! not sure where i will have to insert that 'heartbeat' flag? my commandline looks like this:
thanks again for your supportCode: Select all
"/Applications/OctaneRender Daemon 3.08.1.app/Contents/MacOS/octane_slave_daemon" --slave "/Applications/OctaneRender Daemon 3.08.1.app/Contents/MacOS/octane_slave" --alias "Supercat" --port 48000 --out-of-core 8g; exit
markus
viewtopic.php?p=339955#p339955
and add the log flagheartbeat
to the slave flags file (octane_log_flags.txt
). I have attached a modified archive below.
If you can't get the logging working, I will come back to you tomorrow morning as early as possible.
OctaneRender™ Standalone 3.08.1
You do not have the required permissions to view the files attached to this post.
Specs: Apple MacBook Pro M1 max 64GB 2TB, MacOS 12.5 / MacPro 5,1 with NVIDIA GeForce GTX 1080 8G, MacOS 10.13.6 / Mac Pro 5,1 with AMD RX5700 8G, MacOS 12.3.1 / HP Z600 with NVIDIA 3060 RTX 12G, Windows 10 pro + Netstor GPU box, 4 x NVIDIA GTX 980ti 6G.
Code: Select all
Last login: Wed Jun 27 08:52:04 on ttys000
"/Applications/OctaneRender Daemon 3.08.1.app/Contents/MacOS/octane_slave_daemon" --slave "/Applications/OctaneRender Daemon 3.08.1.app/Contents/MacOS/octane_slave" --alias "Nates-Mac-Pro" --port 48000; exit
Nates-Mac-Pro:~ natemcgraw$ "/Applications/OctaneRender Daemon 3.08.1.app/Contents/MacOS/octane_slave_daemon" --slave "/Applications/OctaneRender Daemon 3.08.1.app/Contents/MacOS/octane_slave" --alias "Nates-Mac-Pro" --port 48000; exit
[13:52:05.122] Started logging on 27.06.18 08:52:05
[13:52:05.122]
[13:52:05.122] Flag Type Value Prefix Description
[13:52:05.122] ------------------------- ------- ---------- ------------ --------------------------------------------------------------------------------
[13:52:05.122] daemon release 1 "daemon" Enables general logging of the slave daemon.
[13:52:05.122] daemonDetail release 1 "daemon" Enables detailed logging of the slave daemon.
[13:52:05.122] logAppendToFile release 1 "" If enabled, logging will be to the already existing log file appended
[13:52:05.122] logFlags release 1 "" If enabled, all log flags will be dumped during startup
[13:52:05.122] logFlushTime release 100 "" The time (ms) after which the log is flushed. If set to 0, the timer based flush will be disabled.
[13:52:05.122] logForceFlush release 0 "" If enabled, every logged line will be flushed
[13:52:05.122] logJuce release 0 "JUCE" If enabled, JUCE logging will go through the logging framework of Octane
[13:52:05.123] logPrefix release 1 "" If enabled, each log line will have a prefix
[13:52:05.123] logTime release 1 "" If enabled, each log line will have a timestamp
[13:52:05.123] logToDebugConsole release 0 "" If enabled, logging will go to the debug console (Windows only)
[13:52:05.123] logToFile release 1 "" If enabled, logging will go to a log file
[13:52:05.123] logToStdout release 1 "" If enabled, logging will go to standard out (when set to 1) or standard error (when set to 2)
[13:52:05.123]
[13:52:05.123] daemon : Command line:
[13:52:05.123] daemon : --slave "/Applications/OctaneRender Daemon 3.08.1.app/Contents/MacOS/octane_slave" --alias Nates-Mac-Pro --port 48000
[13:52:05.123] Launching slave to gather information:
[13:52:05.123]
[13:52:05.123] daemon : using free port 48001 to communicate with slave
[13:52:05.401] Started logging on 27.06.18 08:52:05
[13:52:05.401]
[13:52:05.401] Flag Type Value Prefix Description
[13:52:05.401] ------------------------- ------- ---------- ------------ --------------------------------------------------------------------------------
[13:52:05.401] activation release 0 "activation" Release log for the activation manager
[13:52:05.401] apiChangeMgr release 0 "apiChngMgr" Logs ApiChangeManager::update().
[13:52:05.401] apiForceConnectChecks release 0 "apiChecks" Enforces validation of node-pin connections.
[13:52:05.401] apiNodeSystem release 0 "apiNodeSys" Logs all write operations in the API node system.
[13:52:05.401] apiProject release 0 "apiProject" Logs all important calls into ApiProjectManager.
[13:52:05.401] apiRender release 0 "apiRender" Logs all important calls into ApiRenderEngine.
[13:52:05.401] appShutdown release 0 "appShutdow" Dumps information about application shutdown
[13:52:05.401] cacheLoad release 0 "cacheLoad" Dumps information about the data cache load
[13:52:05.401] cudaAlwaysSync release 0 "cuda" Enables synchronization after each kernel call
[13:52:05.401] cudaProf release 0 "cudaProf" Enables general logging of CUDA profiling.
[13:52:05.401] cudaProfDetail release 0 "cudaProf" Enables logging of single CUDA activity records.
[13:52:05.401] daemon release 1 "network" Not used.
[13:52:05.401] dumpAlembicObjects release 0 "AlembicObj" Dump data about exported Alembic objects
[13:52:05.401] heartbeat release 1 "network" Enables logging of the network rendering heartbeat.
[13:52:05.401] imageSaving release 0 "imageSavin" Logs saving of render images.
[13:52:05.401] licenseUsage release 0 "licUsage" Release logging flag for inter-process license usage
[13:52:05.401] logAppendToFile release 1 "" If enabled, logging will be to the already existing log file appended
[13:52:05.401] logFlags release 1 "" If enabled, all log flags will be dumped during startup
[13:52:05.401] logFlushTime release 100 "" The time (ms) after which the log is flushed. If set to 0, the timer based flush will be disabled.
[13:52:05.401] logForceFlush release 0 "" If enabled, every logged line will be flushed
[13:52:05.401] logJuce release 0 "JUCE" If enabled, JUCE logging will go through the logging framework of Octane
[13:52:05.401] logPrefix release 1 "" If enabled, each log line will have a prefix
[13:52:05.401] logTime release 1 "" If enabled, each log line will have a timestamp
[13:52:05.401] logToDebugConsole release 0 "" If enabled, logging will go to the debug console (Windows only)
[13:52:05.401] logToFile release 1 "" If enabled, logging will go to a log file
[13:52:05.401] logToStdout release 1 "" If enabled, logging will go to standard out (when set to 1) or standard error (when set to 2)
[13:52:05.401] makeAllInspectable release 0 "geometryAr" If enabled, all geometry archives become inspectable. Should only be used if you know what you are doing.
[13:52:05.401] meshValidation release 0 "validation" Logs errors of the mesh data validation
[13:52:05.401] moduleData release 0 "moduleData" Logging for the loading/storage of module data
[13:52:05.402] moduleLoader release 0 "loader" Logging for module loading and unloading.
[13:52:05.402] netRender release 1 "netRender" General logging of network rendering
[13:52:05.402] rdStart release 0 "cuda" Logs information about the creation and start up of render devices and threads.
[13:52:05.402] subdivider release 0 "subdivider" logs info about the subdivision process
[13:52:05.402]
[13:52:05.402] OctaneRender 3.08.1 (3080100)
[13:52:05.402]
[13:52:05.489] network : Starting heartbeat with daemon listening on port 48001
[13:52:05.489] daemon : slave connection created
[13:52:05.489] daemon : slave connection established
[13:52:05.489] daemon : started to monitor slave
[13:52:05.489] network : Connected to daemon
[13:52:05.504] network : Sending heartbeat message 'AUTHENTICATING'
[13:52:05.504] daemon : received heartbeat message 'AUTHENTICATING' from slave
[13:52:08.456] network : Adding sub-system 0x7fb2138c8210 to heartbeat...
[13:52:08.713] Launching net render slave (3080100) with master 0.0.0.0:0
[13:52:08.713] netRender : starting net render slave
[13:52:08.713] network : Adding sub-system 0x7fb214817e00 to heartbeat...
[13:52:08.713] network : Sending heartbeat message 'SLAVE_INFO'
[13:52:08.713] netRender : stopping net render slave ...
[13:52:08.714] daemon : received heartbeat message 'SLAVE_INFO' from slave
[13:52:08.714] daemon : slave has version 3080100 and provides 1 GPU(s)
[13:52:08.715] network : Removing sub-system 0x7fb214817e00 to heartbeat...
[13:52:08.722] network : Removing sub-system 0x7fb2138c8210 to heartbeat...
[13:52:08.722] network : Stopping heartbeat
[13:52:08.723] daemon : slave connection closed
[13:52:08.723] daemon : stopped monitoring slave
[13:52:12.776] detected that slave is hanging after initialization -> stopping slave
[13:52:12.777] ERROR : failed to send stop message to slave
[13:52:14.253]
[13:52:14.253] Stopped logging on 27.06.18 08:52:14
[13:52:14.253]
[13:52:14.379]
[13:52:14.379] Stopped logging on 27.06.18 08:52:14
[13:52:14.379]
logout
Saving session...
...copying shared history...
...saving history...truncating history files...
...completed.
[Process completed]
You do not have the required permissions to view the files attached to this post.
______________
Nate Mac - Chicago, IL
Cinema 4D R26.107 | Octane 2021.1.6
Main: AMD ThreadRipper | Win10 | 32-core | 128GB | RTX A5000 & GTX 2070 SUPER
Slaves: 4x GTX 980 Ti | GTX 1080 | GTX 1070 | GTX 1070Ti
Nate Mac - Chicago, IL
Cinema 4D R26.107 | Octane 2021.1.6
Main: AMD ThreadRipper | Win10 | 32-core | 128GB | RTX A5000 & GTX 2070 SUPER
Slaves: 4x GTX 980 Ti | GTX 1080 | GTX 1070 | GTX 1070Ti
Thank you both for running the test. It seems that the net render slave hangs during shutdown. I'm trying now to figure out where that might be, but if I can't I will make you a special build with more logging to check at which step it exactly hangs and then come back to you. Thank you for your help.
EDIT: Ok, looking at the logs again, it's a bit different: The slave actually shuts down. It just takes a long time (~8.5s), and the daemon should wait 12s but waits only 4s due to a bug. I don't know why that problem doesn't occur in 3.08 and I also don't know which part takes so long to shut down. -> I will sort out the incorrect wait times in the daemon and add a lot of logging to try to find out what's taking so long and send you the build via PM.
EDIT: Ok, looking at the logs again, it's a bit different: The slave actually shuts down. It just takes a long time (~8.5s), and the daemon should wait 12s but waits only 4s due to a bug. I don't know why that problem doesn't occur in 3.08 and I also don't know which part takes so long to shut down. -> I will sort out the incorrect wait times in the daemon and add a lot of logging to try to find out what's taking so long and send you the build via PM.
In theory there is no difference between theory and practice. In practice there is. - Yogi Berra
Thanks for the helpabstrax wrote:Thank you both for running the test. It seems that the net render slave hangs during shutdown. I'm trying now to figure out where that might be, but if I can't I will make you a special build with more logging to check at which step it exactly hangs and then come back to you. Thank you for your help.
EDIT: Ok, looking at the logs again, it's a bit different: The slave actually shuts down. It just takes a long time (~8.5s), and the daemon should wait 12s but waits only 4s due to a bug. I don't know why that problem doesn't occur in 3.08 and I also don't know which part takes so long to shut down. -> I will sort out the incorrect wait times in the daemon and add a lot of logging to try to find out what's taking so long and send you the build via PM.
______________
Nate Mac - Chicago, IL
Cinema 4D R26.107 | Octane 2021.1.6
Main: AMD ThreadRipper | Win10 | 32-core | 128GB | RTX A5000 & GTX 2070 SUPER
Slaves: 4x GTX 980 Ti | GTX 1080 | GTX 1070 | GTX 1070Ti
Nate Mac - Chicago, IL
Cinema 4D R26.107 | Octane 2021.1.6
Main: AMD ThreadRipper | Win10 | 32-core | 128GB | RTX A5000 & GTX 2070 SUPER
Slaves: 4x GTX 980 Ti | GTX 1080 | GTX 1070 | GTX 1070Ti
.
Another week another broken promise.
Another week another broken promise.
Goldorak wrote:We ran into several unforeseen QA issues with the offline mode system which pushed this out further than we hoped - but sorted that out last week. We'll post in this thread with a precise update once we have finished QA process by EOW.Goldorak wrote:Sorry for the inconvenience - while this is otherwise a feature complete release, we should have also noted that the offline system for 3.08+ is still at least a week or more out as we compete testing on a new backend for this. If anyone requires offline mode sooner, I’m recommending staying on 3.07 for just a little longer. Please contact help desk if there is a hard and immediate requirement for offline mode with 3.08.
C4D 2025 | Win10
- DigiChuckle
- Posts: 25
- Joined: Sun Aug 02, 2015 9:24 pm
- Location: Amsterdam
- Contact:
I noticed the alpha channel was aliased while trying to do a back-to-beauty composite in Fusion. This is bad!
The alpha channel in exr will not be anti-aliased whether or not you multiply the color channels with it (by checking the pre-multiplied alpha checkbox in the Imager). This is true for all beauty passes I tried. Not true for info passes.
The work-around while they fix this is to take the alpha from an info pass and transfer that over to the beauty (in a compositing package).
Otoy, is there a way to switch between color and alpha in SA viewport?
The alpha channel in exr will not be anti-aliased whether or not you multiply the color channels with it (by checking the pre-multiplied alpha checkbox in the Imager). This is true for all beauty passes I tried. Not true for info passes.
The work-around while they fix this is to take the alpha from an info pass and transfer that over to the beauty (in a compositing package).
Otoy, is there a way to switch between color and alpha in SA viewport?
You do not have the required permissions to view the files attached to this post.
Win10 Pro | Z620 | 2x 8 core Xeon 2.2 Ghz | 64 GB ram | Quadro K4000 + GTX 1060
Is it possible that you enabled the option disable partial alpha? This should usually be disabled unless you want to compose a render layer on top of the residual layers which have been rendered separately and need this option being enabled (see viewtopic.php?p=269484#p269484).DigiChuckle wrote:I noticed the alpha channel was aliased while trying to do a back-to-beauty composite in Fusion. This is bad!
The alpha channel in exr will not be anti-aliased whether or not you multiply the color channels with it (by checking the pre-multiplied alpha checkbox in the Imager). This is true for all beauty passes I tried. Not true for info passes.
Regarding the pre-multiplied / straight alpha: Octane pre-multiplies alpha if the option is enabled in the imager node. By default OpenEXR image should be saved with this option enabled.
Currently that's not possible in the standalone.The work-around while they fix this is to take the alpha from an info pass and transfer that over to the beauty (in a compositing package).
Otoy, is there a way to switch between color and alpha in SA viewport?
In theory there is no difference between theory and practice. In practice there is. - Yogi Berra
- DigiChuckle
- Posts: 25
- Joined: Sun Aug 02, 2015 9:24 pm
- Location: Amsterdam
- Contact:
Hi abstrax, no, Disable partial alpha is not enabled.
The combined beauty is anti-aliased. However, individual beauty passes are not. Those info passes which have alpha, are anti-aliased.
I don't pre-multiply the alpha because I don't want to have to divide it before color ocrrection in Fusion. Straight alpha + color correction + alpha multiply = no issues. The aliasing however, is a nuisance since I have to manually replace the aliased alpha channel with one which is anti-aliased. Easy for the alpha, indeed, but impossible for the RGB of passes.
According to me alpha multiplication and channel aliasing are 2 unrelated subjects.
The combined beauty is anti-aliased. However, individual beauty passes are not. Those info passes which have alpha, are anti-aliased.
I don't pre-multiply the alpha because I don't want to have to divide it before color ocrrection in Fusion. Straight alpha + color correction + alpha multiply = no issues. The aliasing however, is a nuisance since I have to manually replace the aliased alpha channel with one which is anti-aliased. Easy for the alpha, indeed, but impossible for the RGB of passes.
According to me alpha multiplication and channel aliasing are 2 unrelated subjects.
Win10 Pro | Z620 | 2x 8 core Xeon 2.2 Ghz | 64 GB ram | Quadro K4000 + GTX 1060
- DigiChuckle
- Posts: 25
- Joined: Sun Aug 02, 2015 9:24 pm
- Location: Amsterdam
- Contact:
Hi again! So, I tried four SA versions. The aliasing of beauty passes is a problem in all versions above 3.07 stable. In V4, too. So, what's changed vis-a-vi anti-aliasing since 3.07?
Win10 Pro | Z620 | 2x 8 core Xeon 2.2 Ghz | 64 GB ram | Quadro K4000 + GTX 1060
Nothing that I am aware of. Could you send me an example scene and a description of how you compose in Fusion so I can understand your problem better? Because I couldn't see any issues, if the imager settings are correct.DigiChuckle wrote:Hi again! So, I tried four SA versions. The aliasing of beauty passes is a problem in all versions above 3.07 stable. In V4, too. So, what's changed vis-a-vi anti-aliasing since 3.07?
In theory there is no difference between theory and practice. In practice there is. - Yogi Berra