[vdo-devel] can vdo fix corruption by itself?

[d tbsky]

Sweet Tea Dorminy <sweettea at redhat.com>
>
> > >     Device mapper status: 0 64424509440 vdo /dev/drbd1 normal -
> > > opening online 472234581 2684272631
> > It will change to normal - online online when "index opening" finishes.
> > >     Index sparse: enabled
> > >     Index status: opening
> >
> > Device is online but index is not. I think that until index finish opening deduplication won't work for new writes.
>
> Precisely so. VDO metadata and the index will recover after unclean
> shutdown; VDO recovery is captured in "recovery percent" which is N/A
> when VDO is fully recovered. Index recovery isn't captured in any
> statistic, but as Lukasz said, 'opening' indicates it's recovering,
> and deduplication won't work until it has finished opening ('normal'
> status)
>
     it ran over 24 hours but still can not make it.  I check the IO
speed, sometimes about 5-10MB/sec, sometimes about 20-30MB/sec.
I have 1.9TB data(after compress/dedupe) at vdo. I wonder if vdo need
to  read all the 1.9TB data to finish the job? with the current speed
it may need another day.
but I am worried it will need more time when data grow large.