Edge tf not ready

[HiranmayeeKadali]

Hii I'm trying to running the experiment simple and I get a persistent 'edge tf not ready'. Will it take time to start the freD instance or should I change anything to run the experiment.

[pfandzelter]

Hi, can you provide the full output of the script? Maybe it encounters an error along the way

[HiranmayeeKadali]

preparation.txt
4_movingaverage-10-1-300-10-1-25-100-1-10-cloud-2.txt
Here it is

[pfandzelter]

Hi, it seems that the error is with your version of OpenSSL. Can you provide the output of openssl version on your computer?

[HiranmayeeKadali]

Hii Here it is,
OpenSSL 3.0.2 15 Mar 2022 (Library: OpenSSL 3.0.2 15 Mar 2022)

[pfandzelter]

Ok, I made the necessary changes to read the key file that is produced in newer versions of OpenSSL (PKCS#8 instead of PKCS#1). On macOS, OpenSSL 1.1.1 seems to be the default. You may need to recompile enoki for this, which should be part of the prep-simple.sh script. Make sure you have Go installed on your machine

[HiranmayeeKadali]

Thanks for the support and update. Will update if I face any more issues.

[HiranmayeeKadali]

Hii Again! I'm successful with the "simple" experiment. But for befaas and replication experiment, i get this
preparation.txt
Can you suggest if i'm missing something or should i change anything?:)

[pfandzelter]

Looks like Terraform is deploying to the faasm-422113 project but enoki-422018 is used for connecting the machine. Not sure what is causing this, looks like some configuration on your machine

[HiranmayeeKadali]

Hii @pfandzelter I have a small question, I'm trying to run the replication and scale experiments.
For replication experiment, I can create a client instance because of the following error, "1mError waiting for instance to create: Quota 'N2_CPUS' exceeded. Limit: 8.0" Can you advice any other to run this, I'm trying to contact the GCP service to increase the quota but no response yet.
I can't find the prep_scale.sh file for scale experiment.
I'm trying to work with your work for my thesis work, your advice will be quite helpful.

[pfandzelter]

Hi, regarding the quota you can a) have GCP update your quota to 10 CPUs, b) use other cloud machine types (see variables.tf) so that the experiment is below this quota, or c) use your own servers if you have access to any, but that will require changing the setup scripts accordingly.

There is no prep_scale.sh as that uses the prep.sh script to prepare, see experiment-scale.py

[HiranmayeeKadali]

Hi, Thankyou so much for the advice, but I get file not found error for prep.sh.

ekadsre@ct-blade43:~/enoki$ ./experiment-scale.py
Preparing...
Output: output/output-scale/preparation.txt
Traceback (most recent call last):
File "/home/ekadsre/enoki/./experiment-scale.py", line 183, in
subprocess.run(["./prep.sh"], stdout=of, stderr=of)
File "/usr/lib/python3.10/subprocess.py", line 503, in run
with Popen(*popenargs, **kwargs) as process:
File "/usr/lib/python3.10/subprocess.py", line 971, in init
self._execute_child(args, executable, preexec_fn, close_fds,
File "/usr/lib/python3.10/subprocess.py", line 1863, in _execute_child
raise child_exception_type(errno_num, err_msg, err_filename)
FileNotFoundError: [Errno 2] No such file or directory: './prep.sh'

[pfandzelter]

I see, that should probably be prep-simple.sh. I fixed this now

[kadalihiranmayee]

Hii @pfandzelter! I hope you find this message well. Sorry for disturbing you again, I'm working on your paper for my thesis and need to run the experiment. I'm having some issue with the the Docker image git.tu-berlin.de:5000/mcc-fred/fred/fred:v0.2.18 and git.tu-berlin.de:5000/mcc-fred/fred/alexandra:v0.2.18 cannot be found in the specified registry. I saw they just updated and I'm trying to update the image as well, but can you please look at the document and send me suggested updates as well please, it would be very helpful!
preparation.txt
4_movingaverage-10-1-300-10-1-25-100-1-10-cloud-2.txt

[pfandzelter]

I think you have the right idea, it might be easiest to update the image to "v0.2"