On Wed, Feb 26, 2020 at 10:15 PM Kevin Fenzi <kevin@xxxxxxxxx> wrote: > > On Wed, Feb 26, 2020 at 07:18:53AM +0100, Jan Kaluza wrote: > > Hi, > > > > I'm considering migration of ODCS > > (https://odcs.fedoraproject.org/api/1/composes/) to Openshift in order > > to make it more scalable and also to make it closer to internal ODCS > > deployment which I would like to also move to Openshift in the future. > > I have some questions related to this: > > > > 1) ODCS needs /mnt/koji mounted in read-only mode. Based on my > > information, it should be possible to mount /mnt/koji in read-only > > mode in Openshift. Do you foresee any issues with this? > > The biggest problem with it is that openshift wont let multiple pods use > the same pv, so you have to make N of them confusingly, but I guess > thats not a blocker. > > > 2) ODCS needs /srv/odcs NFS storage ("fedora_odcs") mounted in > > read-write mode. Is it possible and safe to mount this storage in > > Openshift in read-write? > > Yeah, should be. But note only in one pod. Thanks for your response Kevin. Everything is clear except of this ^ one note. I thought I would have multiple backends (pods) running in Openshift and each of them would be able to write to /srv/odcs. Why is it possible to have write access only for single pod? Maybe I could go without /srv/odcs NFS storage completely and just use openshift internal volumes (If it's also possible to create them in Fedora Openshift). But during the devconf, I've been discussing the possibility with Mohan and Tomas Hrcka to create some composes for them using ODCS and I think they will in the end need these composes to be stored on some NFS storage, so they will need a way to access them from outside of Openshift. I thought using /srv/odcs would be ideal for this. Regards, Jan Kaluza > > 3) If I understand it correctly, I will have to use official Fedora > > Openshift (it means *no* comunnishift) even for my initial tests, > > because communishift does not allow me to mount these volumes. Is it > > Correct, you cannot mount any of those volumes there. > You can however use clean new volumes and populate them with whatever > test data you like. So, it may still be wise to do initial work in > communishift and then move on to internal clusters once things are > working? But you are doing the work, so it's up to you of course. :) > > > possible to ask for new Fedora Openshift project for this use-case? > > Well, the way the project exists is via a playbook/configuation in > ansible, so you can come up with a patch and send it for review? > Look at playbooks/openshift-apps and roles/openshift-apps for examples. > > > 4) Do you foresee any issues I might be missing? > > I would think it would work. > > kevin > _______________________________________________ > infrastructure mailing list -- infrastructure@xxxxxxxxxxxxxxxxxxxxxxx > To unsubscribe send an email to infrastructure-leave@xxxxxxxxxxxxxxxxxxxxxxx > Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ > List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines > List Archives: https://lists.fedoraproject.org/archives/list/infrastructure@xxxxxxxxxxxxxxxxxxxxxxx _______________________________________________ infrastructure mailing list -- infrastructure@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe send an email to infrastructure-leave@xxxxxxxxxxxxxxxxxxxxxxx Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/infrastructure@xxxxxxxxxxxxxxxxxxxxxxx
