Closed GoogleCodeExporter closed 8 years ago
Attempted fix in r211 by modifying the queueManager.
We now use a microsecond-based instead of second-based timestamp/time as a part
of our entry_time for key uniqueness in the queue_tickets table and sleep a
random number of microseconds between 100th of a second and a whole second
between each attempt.
In addition it now correctly throws an exception after the 5th attempt if
unsuccessful.
This should both reduce the likelihood for colliding attempts to obtain a
ticket in the GLOBAL queue as well as ensure we correctly catch cases where we
exhaust our retries doing so.
Original comment by lachlan....@gmail.com
on 27 Jan 2012 at 12:15
[deleted comment]
Appears this fix does not address the issue. I realised there is a much simpler
way I can handle the queue mechanism that will avoid the need for this stuff.
Will try to push a fix to trunk today.
Original comment by lachlan....@gmail.com
on 27 Jan 2012 at 6:02
Confirmed fixed in rev 212
Original comment by lachlan....@gmail.com
on 28 Jan 2012 at 2:08
Original issue reported on code.google.com by
lachlan....@gmail.com
on 27 Jan 2012 at 12:10