mirrors/ArchiveBox
1
0
Fork 0
mirror of https://github.com/ArchiveBox/ArchiveBox.git synced 2025-05-18 17:14:39 -04:00
Code Issues Projects Releases Packages Wiki Activity Actions

better actor atomic claim

Browse source
This commit is contained in:
Nick Sweeting 2024-11-02 19:54:25 -07:00
parent 9b24fe7390
commit 2337f874ad
No known key found for this signature in database
2 changed files with 215 additions and 159 deletions
Download patch file Download diff file Expand all files Collapse all files

10
archivebox/actors/orchestrator.py
View file

@ -191,22 +191,26 @@ class FaviconActor(ActorType[ArchiveResult]):
# )
return cls.get_random(
model=ArchiveResult,
where='status = "failed"',
where='status = "failed" AND extractor = "favicon"',
set='status = "queued"',
choose_from_top=cls.get_queue().count(),
choose_from_top=50,
)
def tick(self, obj: ArchiveResult):
print(f'[grey53]{self}.tick({obj.abid or obj.id}) remaining:[/grey53]', self.get_queue().count())
print(f'[grey53]{self}.tick({obj.abid or obj.id}, status={obj.status}) remaining:[/grey53]', self.get_queue().count())
updated = ArchiveResult.objects.filter(id=obj.id, status='started').update(status='success') == 1
if not updated:
raise Exception(f'Failed to update {obj.abid or obj.id}, interrupted by another actor writing to the same object')
# obj.refresh_from_db()
obj.status = 'success'
def lock(self, obj: ArchiveResult) -> bool:
"""As an alternative to self.get_next_atomic(), we can use select_for_update() or manually update a semaphore field here"""
locked = ArchiveResult.objects.filter(id=obj.id, status='queued').update(status='started') == 1
if locked:
# obj.refresh_from_db()
obj.status = 'started'
# print(f'FaviconActor[{self.pid}] lock({obj.id}) 🔒')
pass
else: