<sergiusens> /dev/sr0 is instantly there and cloud-init doesn't choke
<ogra_> ok
<rsalveti> sergiusens: was the module loaded?
<ogra_> obviously ... if he didnt get the error
<rsalveti> guess the kernel tried to read the block device and then loaded the correct module to handle that
* Letozaf_ (~Letozaf@ubuntu/member/letozaf) has joined #snappy
<ogra_> well, mount -t iso9660 should omit any probing
<rsalveti> in the broken case I don't think it can even read the block device
<ogra_> and just blindly load isofs
<rsalveti> that's true 
<ogra_> well, you did the test yourself on your snappy ... mounting /dev/foo doesnt exist either
* deker has quit (Quit: Page closed)
<rsalveti> but in the azure case it was there, just not really there 
<rsalveti> so not sure if that would case any other weird side effect 
* kgunn has quit (Ping timeout: 246 seconds)
<ogra_> well, lets try the loop/wait logic and see
* ezobn has quit (Ping timeout: 240 seconds)
<rsalveti> yeah, in any case, having the retry logic seems a good thing to do anyway
<ogra_> yup
<sergiusens> sorry, intel driver crash
<rsalveti> utlemming: the udev rule could trigger something, but cloud init would still try to read it
* sergiusens missed some bits and pieces