[Xen-API] Some interesting observations of VHD coalesing

[George Shuklin <george.shuklin@xxxxxxxxx>]

Goog day.

Here some observations of XCP 1.1 behavior in case of VHD-tree in SR. In 
my case it was LVMoISCSI:
1) If any host is disabled, VHD is not removed from LMV because xapi ask 
every host in pool about vdi usage before it removing. If host disabled, 
no request been send.
2) After issuing xe sr-scan operations continue in background mode even 
after scan operation is completed.
3) If any host due any reason do have troubles with SR (f.e. freezed LVM 
or something like that) it will prevents any coalese or lvremove 
operations on that LVM.
4) If some host is not responding (f.e. is off due hardware maintenance) 
all background operations with SR will stops.
5) If any VHD is chain (even not used) is broken, all chain is skipped 
and will never be coalesed (still not sure about LV removing after 
removing last VDI in chain).
That behavior very differ from XCP 0.5, which could remove LV from LVM 
is other hosts are unresponsive.


_______________________________________________
Xen-api mailing list
Xen-api@xxxxxxxxxxxxx
http://lists.xen.org/cgi-bin/mailman/listinfo/xen-api