You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hello,
I've deployed ceph using kolla (https://wiki.openstack.org/wiki/Kolla).
The main technical decision of kolla is to use docker for everything and so ceph is deployed "dockerized".
As such, using inkscope with has (very) few issues:
sysprobe needs fsid and use ceph with ceph.confto retrieve it. As there's no ceph in the base os, it's difficult to get it (could use docker exec ceph_* ceph... but it would be complicated to know how to start ceph I think). Perhaps we can add a fsid options in inkscope.confbecause ceph command is used only for that in sysprobe?
cephprobe deeply use ceph commands so proposed solution above would not work. I've then created a docker to embed cephprobe (docker created use kolla based docker image, so I can give the Dockerfile but it will work only on a kolla based deployment. We can also work on a "standalone" docker if you see interest). The issue I have is that cephprobe.py is a daemon and docker needs foreground srcipts. So I've changed the script (the fast way, may be done a lot better) to make it a foreground script. I think that it would be better to rely on OS command (instead of python tools) to do the daemon part as it make it a lot easier to use it.
Except these little issues (the last one may be the trickier), everything works well!
The text was updated successfully, but these errors were encountered:
Hello,
I've deployed ceph using kolla (https://wiki.openstack.org/wiki/Kolla).
The main technical decision of kolla is to use docker for everything and so ceph is deployed "dockerized".
As such, using inkscope with has (very) few issues:
-d
catch in getopt (see allow "-d" as options for ceph-* command #73)fsid
and useceph
withceph.conf
to retrieve it. As there's noceph
in the base os, it's difficult to get it (could usedocker exec ceph_* ceph...
but it would be complicated to know how to start ceph I think). Perhaps we can add afsid
options ininkscope.conf
because ceph command is used only for that in sysprobe?cephprobe.py
is a daemon and docker needs foreground srcipts. So I've changed the script (the fast way, may be done a lot better) to make it a foreground script. I think that it would be better to rely on OS command (instead of python tools) to do the daemon part as it make it a lot easier to use it.Except these little issues (the last one may be the trickier), everything works well!
The text was updated successfully, but these errors were encountered: