Ticket #583 (new defect)

Opened 11 years ago

Last modified 11 years ago

pyhesiodfs should know how to fallback to lower-priority lockers

Reported by: geofft Owned by:
Priority: normal Milestone: The Distant Future
Component: -- Keywords:
Cc: Fixed in version:
Upstream bug:


# FIXME check if the first locker is valid
if len(filsys.filsys) >= 1:
    pointers = filsys.filsys
    pointer = pointers[0]

Change History

comment:1 Changed 11 years ago by geofft

One issue is that open() and the like can block indefinitely if AFS is down. So we should spawn a new thread to open or stat the directory, and fall back if the thread does not terminate within a second or two.

But I have some concerns about that:

  • Can we terminate the thread, if it ends up unkillably in kernel space? Is it fine to leak these threads until AFS resolves itself?
  • Can threading with Python not suck?
  • Can threading with FUSE not suck?

comment:2 Changed 11 years ago by xavid

I would hope that CPython releases the GIL before calling open(). If my optimism is merited, Python threading won't be a problem here.

I've used threading with the C fuse interface before and didn't need to do anything terribly weird.

Note: See TracTickets for help on using tickets.