Closed pibion closed 2 years ago
I did check and there are no other instances running (we get a different error when there aren't enough resources, so this is probably a moot point).
The cluster has been up for more than a year and it's really nice that kubernetes has a security feature 🤔 where the internal certificates expire after a year... So I will need some time to figure out how to properly renew the certificates. I will update here once I'm done.
Certificate renewal failed, and by mistake a couple of months ago I overwrote the kubespray cluster management files on my machine, won't happen again, I will now use separate directory trees. So I need to redeploy. Then I can restore files from backup.
Before redeploying, I asked the Jetstream team if Jetstream 2 is available. If it is, I would prefer to redeploy directly there.
@pibion Jetstream 2 is in early testing and they can let us in. Jetstream 1 will be operational until summer, but they will soon encourage transition.
data
volume, would you like me to back it up to OSN and then re-create on Jetstream 2? Or I'll just provide an empty one?Yes, let's move to Jetstream 2! If you could re-create the data
volume on Jetstream 2 that would be ideal.
It's okay if that fails, though, we can re-populate if necessary.
ok, I requested the transfer to Jetstream 2, we will keep the allocation on Jetstream 1 for a few weeks, so I'll try to directly copy data from the old volume to the new one.
@zonca any success with the authentication on Jetstream 2?
Yes! Will try deployment in the next days
@zonca that's great news! I had a student contact me today asking about our jetstream instance being down - he uses it often for code development. I was unusually happy to get a complaint since it means people are using the system!
ok, it's working on Jetstream 2.
I deployed it temporarily on https://supercdms.zonca.dev/, I'll move to the old url in the next days.
I haven't imported any backup yet.
we also have a newer version of Jupyterhub (JupyterHub 1.5.0 20220317202953
)
ok, I started the transfer of all the data from the data volume, it will take a few hours.
I think it is better if the users first login to the system, then who wants to have their old data restored, can write me and I can copy the data from the backup into their new volumes.
Excellent, I'll have them ping you here if they'd like their backup restored!
@zonca Hi, Zonca - I would like my home directory restored - my username on jetstream is zkromer.
sure @zkromerUCD , can you please first login to the system so the new volume is created?
ok, I can do that
ok, I started the transfer of all the data from the data volume, it will take a few hours.
completed, I transferred 318 GB, notice that the IP to access the data volume changed, see the repository
@zonca great, I can see the expected files in /cvmfs. There is some weirdness with /cvmfs/data/CDMS/Soudan/DMC_V1-5_PhotoneutronSb/Raw/Raw
, not sure if that's an issue with the setup or with how we created the directories.
not sure, I just copied everything with scp -r
from one volume to the other
ok, I consider this completed, @pibion if there is any other issue please open a dedicated issue.
For restoring user data, let's use https://github.com/det-lab/jupyterhub-deploy-kubernetes-jetstream/issues/64 instead.
@pibion @zkromerUCD we are back at the original URL: https://supercdms.jetstream-cloud.org/
Spawning a jupyter instance gets stuck in "your server is starting up" mode for more than several minutes and then fails.