channeldb+routing+gossiper: add local updates to graph immediately #4964
Conversation
halseth
changed the title
halseth
force-pushed
the
gossip-local-no-batch
branch
2 times, most recently
from
fedac28
to
1b182bf
Compare
batch/interface.go
Outdated
| lazy bool | ||
| } | ||
|
|
||
| // SchedulerOption is a type that can be can supply options to a scheduled |
|
There appears to be an itest failure related to private channels. |
halseth
force-pushed
the
gossip-local-no-batch
branch
2 times, most recently
from
f3d8d99
to
23f7958
Compare
|
The flakes seems to just be the same as those seen in master, not clear whether this PR makes it worse. I think we should wait until we have #4953 or similar merged. |
halseth
force-pushed
the
gossip-local-no-batch
branch
from
23f7958
to
7993d19
Compare
@halseth are they? I'm seeing blanket failures across all the itests for this PR. |
halseth
force-pushed
the
gossip-local-no-batch
branch
from
7993d19
to
cea1b1c
Compare
Yeah, looks like this makes it worse indeed. I still think the root cause is fixed by #4953 though, probably some timing interaction of this PR that makes it more likely to flake. |
halseth
force-pushed
the
gossip-local-no-batch
branch
from
cea1b1c
to
5a3e195
Compare
|
Rebased on #5003 to check travis |
We make the default non-lazy, and will make the incoming gossip requests lazy.
Since the batch interval can potentially be long, adding local updates to the graph could be slow. This would slow down operations like adding our own channel update and announcements during the funding process, and updating edge policies for local channels. Now we instead check whether the update is remote or not, and only for remote updates use the SchedulerOption to lazily add them to the graph.
halseth
force-pushed
the
gossip-local-no-batch
branch
from
5a3e195
to
c9afc93
Compare
|
rebased on master |
| // If this is a remote update, we set the scheduler option to lazily | ||
| // add it to the graph. | ||
| var schedulerOp []batch.SchedulerOption | ||
| if nMsg.isRemote { |
Since the batch interval can potentially be long, adding local updates
to the graph could be slow. This would slow down operations like adding
our own channel update and announcements during the funding process, and
updating edge policies for local channels.
Now we instead check whether the update is remote or not, and use the
SchedulerOption to immediately add local updates to the graph.
Companion PR to #4958