problem with device validation

Hi, Peter, i need your help !

I'm still working on a new redundant setup that allow intelligent fr1 raid in 
all cases of failures : when secondary is rebooted (normal behavior), and when 
primary is rebooted too.

to achieve that, I'm running fr1 degraded array on secondary when primary fails, 
   when primary comes back, it is first fr1-synced on secondary, clean array is 
stopped, and then restarted on primary. This way bitmap is created on secondary 
when primary is down and will track changes during the failure.


I have a problem with nbd device validation before adding it to fr1 array :

to be assembled, nbd device has to be validated. But i haven't found a reliable 
way to force it on client side. enbd-server and ennd-client pairs are running 
both side on different ports.

If I restart server it works, but i have to do this on the other machine via ssh 
and i want to avoid this.

If i send USR1 to enbd-client, sometimes it works, but not all the times...

If I do nothing, it is working sometimes too...

Maybe i missed something ?

here is a copy of my nbdinfo :


Device a:       Open
[a] State:      verify, rw, enabled, validated, show_errs, last error 0, lives 
13, bp 0
[a] Queued:     +-2R/0W curr (check 0R/0W) +62R/31W max
[a] Buffersize: 262144  (sectors=512, blocks=512)
[a] Blocksize:  512     (log=9)
[a] Size:       104857KB
[a] Blocks:     0
[a] Sockets:    2       (*)     (+)
[a] Requested:  194.03K (101K)  (92.1K) 200R/193.8KW    max 128
[a] Despatched: 194.02K (101K)  (92.1K) 192R/193.8KW    md5 0W (0 eq, 0 ne, 0 dn)
[a] Errored:    56      (0)     (0)     0+56
[a] Pending:    2       (0)     (0)     2R/0W+0R/0W
[a] B/s now:    0       (0R+0W)
[a] B/s ave:    1.42M   (1.00KR+1.42MW)
[a] B/s max:    3.60G   (20.5MR+3.58GW)
[a] Spectrum:   7%1     2%8     90%128
[a] Kthreads:   0       (0 waiting/0 running/1 max)
[a] Cthreads:   2       (+)     (+)
[a] Cpids:      2       (19835) (18983)