Open jeremyatourville opened 5 months ago
Hi @jeremyatourville This is surprising. We tried reproducing it, albiet with the latest images, and it worked. Our usual advise would be to drop the database and try again, although it looks like you did create the fresh database directory.
@mikedep333 When you say drop the database - is there a specific command to run? I am getting a message that you can't drop the currently open DB. Or are you saying delete everything under the pgsql folder?
Alternatively, what image versions were you able to make work? My understanding is that the container image tag latest
is sort of a relative reference. It is the latest at the time of download but there may be changes if you download it at a later time.
I also tried to modify my compose file and run pulp-minimal:3.54.0
and redis:7.2.5
This had no effect on the error. It still tells me ERROR: relation "core_artifact" does not exist at character 28
I also have the problem with docker compose -f compose.folders.yml up
on rock linux 9 and Macbook Pro with DockerCE.
While docker compose -f compose.yml up
is ok.
@mikedep333 Anything to add?
@wahello can you post your docker compose file, the result of docker-compose up -d and the log from when you are starting up the containers? (similar to what I posted above)
docker-compose up -d && docker-compose logs -f
I did try making some changes to my docker-compose.yml file. I changed the container tags being used. I now have the following:
postgres:13
redis:7.2.5
pulp-minimal:3.54.0
pulp-web:latest
This made no change. I still get the same error as my original post.
It resolved by making a tmp
directory in bind volume folder pulp_storage
and chmod a+w tmp
.
Maybe, It's about absence of writing permission.
I did try making some changes to my docker-compose.yml file. I changed the container tags being used. I now have the following:
postgres:13 redis:7.2.5 pulp-minimal:3.54.0 pulp-web:latest
This made no change. I still get the same error as my original post.
It resolved by making a tmp directory in bind volume folder pulp_storage
Hmmmm... that's intersting behavior. I would not have expected that but it totally makes sense. In the example I posted, my docker-compose file is using bind mount in Linux. However, when I tried the same compose setup on Windows with docker desktop I chose to use native docker volumes and I still was getting that behavior. So I would have never guessed that. I'm glad to hear it's working for you. I'm going to give it a shot on my Linux machine and see what happens.
So I noticed today that my docker log shows this info:
pulp-migration_service-1 | crypto/fips/fips.c:154: OpenSSL internal error: FATAL FIPS SELFTEST FAILURE
I didn't see that in my earlier log.
Version docker.io/library/postgres:13 docker.io/library/redis:latest pulp/pulp-minimal:latest pulp/pulp-web:latest
Describe the bug
service "migration_service" didn't completed successfully: exit 139
is seen when running docker-compose up -d docker logs show this line -ERROR: relation "core_artifact" does not exist at character 28
To Reproduce Use compose.folders.yaml from git mkdir ./pgsql ./pulp_storage sudo chown 700:700 ./pulp_storage sudo chown 26:26 ./pgsql
Expected behavior DB Migration should complete without issues
Additional context I am running Rocky 8.9 and Docker/Docker-Compose My system is air-gapped and I am pulling images from my private repo.
Folder structure: