This is a plot of the number of blocks within a gdev session.
I was expecting it to be exactly 600 outside of periods with missing blocks but it seems to drop as low as 413 blocks for session 7194.
Seems I have to revise the definition of a session…
This is a plot of the number of blocks within a gdev session.
I was expecting it to be exactly 600 outside of periods with missing blocks but it seems to drop as low as 413 blocks for session 7194.
Seems I have to revise the definition of a session…
If we look at bloc 4125935 (beginning of session 7194), we see:
Some offence and a lot of outgoing authorities.
It seems like a bug to see three times authorityMembers.OutgoingAuthorities ![]()
→ #281
Why? If authorities expire or there is a bug that makes validators crash, the session has <600 blocks, that’s normal.
The graph is even reassuring: it never drops under 2/3.
We have 2356 sessions with 600 blocks which is less than 30%. Are there so many bugs we did not see?
Voilà la répartition des intervalles de temps entre blocs de 4125935 à 4126348 (412 blocs de la session 7194 en 3564 secondes) :
Soit 283×1 + 90×2 + 29×3 + 5×4 + 2×5 + 2×7 = 594 slots sur 600 (j’ai juste oublié le premier et dernier bloc dans mes données).
Le bloc 4,126,014 vient 42 secondes après le précédent.