Am 29.06.21 um 01:19 schrieb Victor Hooi:
Hi,
Yes, I can confirm that file exists on my system:
root@angusmtv-vm01:/etc/sudoers.d# cat ceph-osd-smartctl
## allow ceph-osd (which runs as user ceph) to collect device health
metrics
ceph ALL=NOPASSWD: /usr/sbin/smartctl -x --json=o /dev/*
ceph ALL=NOPASSWD: /usr/sbin/nvme * smart-log-add --json /dev/*
This was a fresh install of Proxmox 7.0 beta, with Ceph, I haven't really
made any changes outside of setting the Ceph keyring, so that ceph-volume
batch would work:
ceph auth get client.bootstrap-osd >
/var/lib/ceph/bootstrap-osd/ceph.keyring
The nvme-cli package was *not* installed:
# dpkg -s nvme-cli
dpkg-query: package 'nvme-cli' is not installed and no information is
available
Use dpkg --info (= dpkg-deb --info) to examine archive files.
I've just installed it - should that make those errors go away?
Yes. I now debugged this a bit further: Ceph uses 'sudo nvme ...' when
executing the command, but if 'nvme' can't be found in the PATH, the
sudoers rule does not seem to apply, because it is for '/usr/sbin/nvme'.
Seems like something with sudo changed too. Given the same sudoers file
and not having 'nvme-cli' installed, for an unprivileged user using
sudo nvme qemu smart-log-add --json /dev/sdc
and Ctrl+D-ing the password prompt results in a notification in PVE 7.0,
but not in PVE 6.4.
Should we install nvme-cli by default then?
It is in Recommends for the ceph package, but when installing via the
API we use 'apt --no-install-recommends' for some reason. Maybe somebody
else knows why?
Out of curiosity - where do you see the hardware status via Ceph
I think you can use
ceph device ls
to get the list of devices and then
ceph device get-health-metrics <DEVICE>
Thanks
Victor
On Mon, Jun 28, 2021 at 5:54 PM Fabian Ebner <f.eb...@proxmox.com> wrote:
Am 25.06.21 um 03:31 schrieb Victor Hooi:
Hi,
I've setup a new Ceph cluster with the Proxmox 7.0 beta
I'm getting notification emails like this:
ceph : a password is required ; PWD=/ ; USER=root ; COMMAND=nvme intel
smart-log-add --json /dev/nvme0n1
Do you know if this is expected behaviour, or what it means? Is this some
kind of new logging in Ceph 16 or Proxmox 7?
Thanks,
Victor
_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Hi,
in addition to what Dominik said: Is nvme-cli installed?
Best Regards,
Fabi
_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
