Signed-off-by: Stefan Priebe <s.priebe@profihost.ag>

Currently when we open a new summary panel

- we display the rrd with default hour timeframe
- then we refresh it
- then we update from timeframe selector

With this patch:
- We display the rrd with timeframe from selector, or if not defined default hour
- We update from timeframe selector value if timeframe && cf are differents from current value
Signed-off-by: Alexandre Derumier <aderumier@odiso.com>

Java archives (.jar) are simple zip files - already compressed.

Signed-off-by: Alexandre Derumier <aderumier@odiso.com>

This can be used for the new async qemu config API.

Only enable timeouts when we really expect data to be send/received.

And use new API entry point move_disk (changed in recent qemu-server)

Move for unused volumes is not implememnted.

Signed-off-by: Alexandre Derumier <aderumier@odiso.com>

People run into problems on upgrades, for example:

 export http_proxy=http://172.xxx.xxx.xxx:8888/
 aptitude update && aptitude full-upgrade

This restarts pveproxy, and after that login was no longer possible.

Another way to reproduce the bug is:

 http_proxy=http://1.2.3.4:8888/ pveproxy --debug

The proxy call is done async, so the original connection can be
closed already. This just adds a reasonable warning message. We
can remove the message later if that turns out to work as expected.