Hi everybody,
via virsh pool-define-as
we created a pool using an NFS storage. The pool could be started and is visible in both UVMM and virsh.
<pool type='netfs'>
<name>isopool</name>
<uuid>22d29f9a-00df-4e8c-9df5-765c6f7c199c</uuid>
<capacity unit='bytes'>5892566810624</capacity>
<allocation unit='bytes'>2482581798912</allocation>
<available unit='bytes'>3409985011712</available>
<source>
<host name='s01'/>
<dir path='/volume1/isopool'/>
<format type='auto'/>
</source>
<target>
<path>/mnt/isopool</path>
<permissions>
<mode>0777</mode>
<owner>111</owner>
<group>122</group>
</permissions>
</target>
</pool>
When we start a guest-system, having an ISO-file mounted from that nfs pool, we got a permission denied message:
Fehler beim Starten der Domain: internal error: process exited while connecting to monitor: 2018-01-16T15:52:13.799517Z qemu-system-x86_64: -drive file=/mnt/isopool/UIU-PE10-Capture-64.iso,format=raw,if=none,id=drive-ide0-0-0,readonly=on: Could not open '/mnt/isopool/UIU-PE10-Capture-64.iso': Permission denied
Traceback (most recent call last):
File "/usr/share/virt-manager/virtManager/asyncjob.py", line 88, in cb_wrapper
callback(asyncjob, *args, **kwargs)
We already changed the ownership to libvirt-qemu user and the permissions file and folder 0777.
The error seems to be a well known behaviour, when selinux is involved. Often we got the suggestion to
setsebool -P virt_use_nfs 1
UCS is not using selinux, right? Because the setsebool command could not be found on our UCS 4.2.
Cheers
Sebastian