[Tarantool-patches] [PATCH 1/2] gc/xlog: delay xlog cleanup until relays are subscribed

[Serge Petrenko]

20.03.2021 01:17, Vladislav Shpilevoy пишет:
> But what I don't understand now - how does it work if the struct
> replica objects create the consumer objects right in replica_set_id()?
>
> Look at relay_subscribe(). If the replica is not anon, then
> 'replica->id != REPLICA_ID_NIL' is true (because there is an assertion).
> It means replica_set_id() was already called. And this means replica->gc
> is already not NULL. Therefore the check "replica->gc == NULL && !replica->anon"
> is never true. Am I missing something?
replica_set_id() only sets a gc consumer for a previously anonymous replica,
which has just registered. This is not needed at all, likely. Because
box_process_register() already creates a gc consumer. And replaces replica's
gc consumer, if any, with this freshly created one later. So this piece of
code in replica_set_id() is extraneous.

It only  makes sense when an anonymous replica was connected to some
instance, then it registers on another instance and the first instance 
receives
anonymous replica's _cluster registration via replication. This is still 
unneeded.
Because once previously anonymous replica resubscribes, it'll get a gc 
consumer
like everyone else does.

This is quite misleading, indeed. It took me a while to understand 
what's happening.
I'll probably return with a patch removing this piece in 
replica_set_id() altogether.

-- 
Serge Petrenko