yeah... still seeing variances from 1m to 10m for the same task. been testing parallel execution for hours.
from the logs, it feels like after git clone, it spend minutes without outputting anything. @<1523701205467926528:profile|AgitatedDove14> Do you know what is the agent suppose to do after git clone ?
I guess a check that all packages is installed ? But then with CLEARML_AGENT_SKIP_PYTHON_ENV_INSTALL=1, what is the agent doing ??
are you on clearml agent 1.8.0?
(im noticing sometimes im just missing logs such as "Running task id.." entirely)
Please refer to here None
The doc need to be a bit clearer: one require a path and not just true/false
im not running in docker mode though - im running a clearml worker in a docker container (and then multiplying the container)
normally when new package need to be install, it shows up in the Console tab
what if the preexisting venv is just the system python? my base image is python:3.10.10 and i just pip install all requirements in that image. Does that not avoid venv still?
it will basically create a new venv inside the container forking the existing preinistalled stuff (i.e. the new venv already has everything the python system has preinstalled)
then it will call "pip install" on all the "installed packages of the Task.
Which should just check everything is there and install nothing
If you set " CLEARML_AGENT_SKIP_PYTHON_ENV_INSTALL=1" it will do checks and just use the existing system python environment as is.
, I can get 50 tasks to run in the same time it takes to run a single one? i cant imagine the apiserver being a noticeable bottleneck.
50 containers on a single machine would be fine if you have enough RAM/CPU, and yes they would run concurrently.
regrading the time itself, again the spinup time of a Task should be negligible.
Pipeline tasks are not meant to be "threads" they are meant as different functions you want to run on different machines,
This means that if your pipeline is just a set of simple functions that require no cpu/gpu or IO, I'm not sure pipeline steps is the right way to go
Does that make sense?
okay that's a similar setup to mine... that's interesting.
much more in line with my expectation.
thank you!
i'll take that design into consideration.
re: CLEARML_AGENT_SKIP_PYTHON_ENV_INSTALL in "docker venv mode" im still not quite sure I understand correctly - since the agent is running in a container, as far as it is concerned it may as well be on bare-metal.
is it just that there's no way for that worker to avoid venv? (i.e. the only way to bypass venv is to use docker-mode?)
in my case using self-hosted and agent inside a docker container:
47:45 : taks foo pulled
[ git clone, pip install, check that all requirements satisfied, and nothing is downloaded]
48:16 : start training
i would love some advice on that though - should I be using services mode + docker and some max # of instances to be spinning up multiple tasks instead?
my thinking was to avoid some of the docker overhead. but i did try this approach previously and found that the container limit wasn't exactly respected.
im not running in docker mode though
hmmm that might be the first issue. it cannot skip venv creation, it can however use a pre-existing venv (but it will change it every time it installs a missing package)
so setting CLEARML_AGENT_SKIP_PYTHON_ENV_INSTALL=1 in non docker mode has no affect
but pretty reliably some proportion of tasks still just take a much longer time. 1m - 10m is a variance i'd really like to understand.
of what task? i'm running lots of them and benchmarking execution times. would you like to see a best case or worst case scenario? (ive kept some experiments for each).
and yeah, in those docs you just linked, "boolean" vars like CLEARML_AGENT_GIT_CLONE_VERBOSE explicitly say true so I ended up trying that pattern. but originally i did try 1. let me go back to that now. thank you.
overall I've seen some improvements in execution time using the suggestions in this thread (tysm!) - the preinstalled libs seem to be helping, though some things are still just unbearably slow (one of my larger pipelines took > 1 h to generate a DAG before even starting...).
I think a proper screenshot of the full log with some information redacted is the way to go. Otherwise we are just guessing in the dark
oh it's there, before running task.
from task pick-up to "git clone" is now ~30s, much better.
though as far as I understand, the recommendation is still to not run workers-in-docker like this:
export CLEARML_AGENT_SKIP_PYTHON_ENV_INSTALL=1
export CLEARML_AGENT_SKIP_PIP_VENV_INSTALL=$(which python)
(and fwiw I have this in my entrypoint.sh )
cat <<EOF > ~/clearml.conf
agent {
vcs_cache {
enabled: true
}
package_manager: {
type: pip,
system_site_packages: true,
}
}
EOF
what if the preexisting venv is just the system python ? my base image is python:3.10.10 and i just pip install all requirements in that image . Does that not avoid venv still?
it's good to know that in theory there's a path forward with almost zero overhead . that's what I want .
is it reasonable to expect that with sufficient workers, I can get 50 tasks to run in the same time it takes to run a single one? i cant imagine the apiserver being a noticeable bottleneck .
ah I see. thank you very much!
trying export CLEARML_AGENT_SKIP_PIP_VENV_INSTALL=$(which python)
but I still see Environment setup completed successfully
(it is printed after Running task id )
it still takes a full 3 minutes between task pulled by worker until Running task id
is this normal? What is happening in these few minutes (besides a git pull / switch)?
from task pick-up to "git clone" is now ~30s, much better.
This is "spent" calling apt update && update install && pip install clearml-agent
if you have those preinstalled it should be quick
though as far as I understand, the recommendation is still to not run workers-in-docker like this:
if you do not want it to install anything and just use existing venv (leaving the venv as is) and if something is missing then so be it, then yes sure that the way to go
@<1523701205467926528:profile|AgitatedDove14> About why we stay on 1.12.2 : None
ha! yup. that was it exactly. I posted about it too None lol
oh yes. Using env until the next message is 2 minutes.
there is almost zero overhead if your docker container alreadyt has everything (including the agent) preinstalled and you set it with CLEARML_AGENT_SKIP_PYTHON_ENV_INSTALL=1
it then should basically just run the code.
would those containers best be started from something in services mode?
Yes as long as the machine has enough cpu/ram
Notice that the services mode will start a second parallel Task after the first one is done setting up the env, if running with CLEARML_AGENT_SKIP_PYTHON_ENV_INSTALL, with containers that have git/python/clearml-agent preinstalled it should be minimal.
or is it possible to get no-overhead with my approach of worker-inside-docker?
No do not do that, see above explanation on why CLEARML_AGENT_SKIP_PYTHON_ENV_INSTALL does not work in docker venv mode
i designed my tasks as different functions, based mostly on what metrics to report and artifacts that are best cached (and how to best leverage comparisons of tasks). they do require cpu, but not a ton.
just report a single Task as multiple "titles" then each title is it's own step, then inside the "title" they have different seriese
is there a way for me to toggle CLEARML's log level?
Try to set the python master logger base logging level
fwiw - i'm starting to wonder if there's a difference between me "resetting the task" vs cloning it.
oooh thank you, i was hoping for some sort of debugging tips like that. will do.
from a speed-of-clearing-a-queue perspective, is a services-mode queue better or worse than having many workers "always up"?
- try with the latest RC
1.8.1rc2
, it feels like after git clone, it spend minutes without outputting anything
yeah that is odd , can you run the agent with --debug (add before the daemon command) , and then at the end of the command add --foreground
Now launch the same task on that queue, you will have a verbose log in the console.
Let us know what you see
i just ran a pipeline that took about 2h (more than half this time was just the DAG), with about a hundred tasks. i'm taking a look at them now to see what the logs show for runtimes.