6 Apr
2021
6 Apr
'21
12:34 a.m.
Makes sense.
My container "just worked™" because it only ever had python3.6 installed.
On Tue, Apr 6, 2021 at 5:40 AM Patrick Donnelly <pdonnell(a)redhat.com> wrote:
Took another look at this: it looks like just removing python3.8 and
rebuilding Ceph is all that was needed to fix this.
On Mon, Apr 5, 2021 at 9:32 AM Patrick Donnelly <pdonnell(a)redhat.com> wrote:
--
Cheers,
Brad
Hi Brad,
Looks like this is because python3.6 is the default on RHEL8 but ceph
depends on 3.8:
pdonnell@vossi04 ~/ceph$ find /usr/ -name \*prettytable\*
/usr/lib/python3.6/site-packages/prettytable-0.7.2-py3.6.egg-info
/usr/lib/python3.6/site-packages/__pycache__/prettytable.cpython-36.pyc
/usr/lib/python3.6/site-packages/__pycache__/prettytable.cpython-36.opt-1.pyc
/usr/lib/python3.6/site-packages/prettytable.py
find: ‘/usr/share/polkit-1/rules.d’: Permission denied
/usr/share/licenses/python3-prettytable
/usr/share/doc/python3-prettytable
find: ‘/usr/share/selinux/targeted/default/active’: Permission denied
find: ‘/usr/libexec/initscripts/legacy-actions/auditd’: Permission denied
pdonnell@vossi04 ~/ceph$ head build/bin/ceph
#!/usr/bin/python3.8
This RHEL has py3.6 and py3.8 installed. But, the prettytable
installed by install-deps.sh is in the python3.6 site-packages. I
wonder how we normally get around this?
On Fri, Apr 2, 2021 at 4:36 PM Brad Hubbard <bhubbard(a)redhat.com> wrote:
--
Patrick Donnelly, Ph.D.
He / Him / His
Principal Software Engineer
Red Hat Sunnyvale, CA
GPG: 19F28A586F808C2402351B93C3301A3E258DD79D
Hey Patrick,
I tried to reproduce this but couldn't but my process may give you
something to compare to so you can work out what is different in your
environment?
Following is the specifics of what I did.
I used this image.
$ podman images --digests registry.access.redhat.com/ubi8
REPOSITORY TAG DIGEST
IMAGE ID CREATED
SIZE
registry.access.redhat.com/ubi8 latest
sha256:aec01a874399c3a3c31033b35761b28aef605504ade888f98f637442cc56b5c0
4199acc83c6a 7 weeks ago 213 MB
Needed this patch to enable a repo in container mode (I'll create a PR
once I do some more testing).
diff --git a/install-deps.sh b/install-deps.sh
index a5e19998f1..368fea017d 100755
--- a/install-deps.sh
+++ b/install-deps.sh
@@ -363,7 +363,7 @@ else
$SUDO dnf config-manager --add-repo
http://apt-mirror.front.sepia.ceph.com/lab-extras/8/
$SUDO dnf config-manager
--setopt=apt-mirror.front.sepia.ceph.com_lab-extras_8_.gpgcheck=0
--save
elif test $ID = rhel -a $MAJOR_VERSION = 8 ; then
- $SUDO subscription-manager repos --enable
"codeready-builder-for-rhel-8-${ARCH}-rpms"
+ $SUDO dnf config-manager --set-enabled
"codeready-builder-for-rhel-8-${ARCH}-rpms"
$SUDO dnf config-manager --add-repo
http://apt-mirror.front.sepia.ceph.com/lab-extras/8/
$SUDO dnf config-manager
--setopt=apt-mirror.front.sepia.ceph.com_lab-extras_8_.gpgcheck=0
--save
fi
# dnf -y install git
Clone the ceph repo and patch and run install-deps.sh and do_cmake.sh.
# ninja -j8 vstart-base
# MON=1 MDS=0 MGR=1 OSD=3 ../src/vstart.sh --memstore -n -x -l -d
--without-dashboard
# bin/ceph -s
*** DEVELOPER MODE: setting PATH, PYTHONPATH and LD_LIBRARY_PATH ***
2021-04-02T23:14:05.232+0000 7f3571593700 -1 WARNING: all dangerous
and experimental features are enabled.
2021-04-02T23:14:05.244+0000 7f3571593700 -1 WARNING: all dangerous
and experimental features are enabled.
cluster:
id: cd13c7aa-ef6f-4416-a624-d3c082ebaec9
health: HEALTH_WARN
3 mgr modules have failed dependencies
...
HTH you to track it down mate and let me know if I can help.
On Sat, Apr 3, 2021 at 3:39 AM Patrick Donnelly <pdonnell(a)redhat.com> wrote:
--
Patrick Donnelly, Ph.D.
He / Him / His
Principal Software Engineer
Red Hat Sunnyvale, CA
GPG: 19F28A586F808C2402351B93C3301A3E258DD79D
I had a clean install done on my dev machine (RHEL 8.3) which required
me to reinstall dependencies for Ceph. I'm getting this error when
setting up a vstart cluster:
pdonnell@vossi04 ~/ceph/build$ bin/ceph status
*** DEVELOPER MODE: setting PATH, PYTHONPATH and LD_LIBRARY_PATH ***
Traceback (most recent call last):
File "bin/ceph", line 153, in <module>
from ceph_daemon import admin_socket, DaemonWatcher, Termsize
File "/home/pdonnell/ceph/src/pybind/ceph_daemon.py", line 24, in
<module>
from prettytable import PrettyTable, HEADER
ModuleNotFoundError: No module named 'prettytable'
pdonnell@vossi04 ~/ceph/build$ python3
Python 3.6.8 (default, Aug 18 2020, 08:33:21)
[GCC 8.3.1 20191121 (Red Hat 8.3.1-5)] on linux
Type "help", "copyright", "credits" or "license"
for more information.
>>> import prettytable
>>>
pdonnell@vossi04 ~/ceph/build$ python2
bash: python2: command not found
The module is available but bin/ceph is claiming not to find it.
Anyone know what's going on? (I did do install-deps.sh without issue
before running vstart.sh)
--
Patrick Donnelly, Ph.D.
He / Him / His
Principal Software Engineer
Red Hat Sunnyvale, CA
GPG: 19F28A586F808C2402351B93C3301A3E258DD79D
_______________________________________________
Dev mailing list -- dev(a)ceph.io
To unsubscribe send an email to dev-leave(a)ceph.io
--
Cheers,
Brad