Hi all,
The current 2.4 field formatting assigns the following:
RS_UNKNOWN="(unknown)"
RS_NODATA="(nodata)"
RS_UNAVAIL="(unavail)"
RS_OFFLINE="(offline)"
IMHO this is too verbose and not really helpful, especially as these
value usually replicate for most fields of the node:
Example:
node2 offline, node3 not vm_capable:
node1# gnt-node list node{1,2,3}
Node DTotal DFree MTotal MNode MFree Pinst Sinst
node1 698.6G 630.5G 32.0G 1.0G 30.0G 8 7
node2 (offline) (offline) (offline) (offline) (offline) 9 4
node3 (unavail) (unavail) (unavail) (unavail) (unavail) 0 0
node3 is down, but not set offline:
node1# gnt-node list node{1,2,3}
Node DTotal DFree MTotal MNode MFree Pinst Sinst
node1 698.6G 630.5G 32.0G 1.0G 30.0G 8 7
node2 698.6G 632.8G 32.0G 1.0G 30.0G 9 4
node3 (nodata) (nodata) (nodata) (nodata) (nodata) 0 0
I propose that we change back to a similar model as in Ganeti 2.3 and
below, where the only special state was '?', except that we extend with
other special chars. My proposal would be:
RS_UNKNOWN="??"
RS_NODATA="?"
RS_UNAVAIL="-"
RS_OFFLINE="*"
After the change:
Node DTotal DFree MTotal MNode MFree Pinst Sinst
node1 698.6G 630.5G 32.0G 1.0G 30.0G 8 7
node2 * * * * * 9 4
node3 - - - - - 0 0
Node DTotal DFree MTotal MNode MFree Pinst Sinst
node1 698.6G 630.5G 32.0G 1.0G 30.0G 8 7
node2 698.6G 632.8G 32.0G 1.0G 30.0G 9 4
node3 ? ? ? ? ? 0 0
The usual node dead is still ?, whereas a node offline is now *. An
unavailable item, which doesn't apply at all to this entity, is -.
Advantages: the display is more readable, and more compact (keeping the
same width no matter node state)
Disadvantages: the fields might seem cryptic, but I believe the special
chars are well enough chosen that it makes sense after a short period of
accommodation.
Thoughts?
iustin
