I noticed that after I suspend/resume, top stops showing CPU usage for processes for a while (a minute or so). This is not a huge issue, but it’s a bit annoying because I use the AppVM for development, testing etc.
The top header still shows significant CPU usage (say, 30%, which is more than once core), but all the processes show 0% CPU usage, so sort by that field does nothing. And I can see the processes running / processing requests, etc.
After a while things fix on their own, and top starts showing everything just like before.
I’m not quite sure what to look for. But dmesg in the AppVM shows this:
[ 4608.766105] Freezing user space processes
[ 4608.767857] Freezing user space processes completed (elapsed 0.001 seconds)
[ 4608.767883] OOM killer disabled.
[ 4608.767893] Freezing remaining freezable tasks
[ 4608.768850] Freezing remaining freezable tasks completed (elapsed 0.000 seconds)
[ 4608.768873] xen:manage: Using freeze/restore/thaw for sleep/wakeup
[ 4608.769189] suspending xenstore...
[ 4608.769557] xen:grant_table: Grant tables using version 1 layout
[ 4608.771035] OOM killer enabled.
[ 4608.771052] Restarting tasks ... done.
I wonder if this is related to “Freezing user space processes” in some way. But I haven’t noticed anything obviously related once it fixes itself …