Hi all,
unfortunately I have some issues with a web2py instance (1.99.4)
only in production: I can't reproduce the problem, but maybe this is
an easy fix.
Traceback:
Traceback (most recent call last):
File "/home/niphlod/production/web2py/gluon/main.py", line 493, in
wsgibase
serve_controller(request, response, session)
File "/home/niphlod/production/web2py/gluon/main.py", line 187, in
serve_controller
environment = build_environment(request, response, session)
File "/home/niphlod/production/web2py/gluon/compileapp.py", line
339, in build_environment
c = environment['cache'] = Cache(request)
File "/home/niphlod/production/web2py/gluon/cache.py", line 376, in
__init__
self.disk = CacheOnDisk(request)
File "/home/niphlod/production/web2py/gluon/cache.py", line 261, in
__init__
os.unlink(self.shelve_name)
OSError: [Errno 2] No such file or directory: '/home/niphlod/
production/web2py/applications/plaza/cache/cache.shelve'
Looking closely to gluon/cache.py, it seems that it can be fixed ...
here's an excerpt from line 255 and on...
except ImportError:
pass # no module _bsddb, ignoring exception now so it
makes a ticket only if used
except:
logger.error('corrupted file %s, will try delete it!'
\
% self.shelve_name)
try:
os.unlink(self.shelve_name)
except IOError:
logger.warn('unable to delete file %s' %
self.shelve_name)
if locker_locked:
portalocker.unlock(locker)
if locker:
locker.close()
I don't know why it happens, but adding an OSError exception to the
IOError on line 262 will fix it.
BTW: I'm sure I don't call cache on disk within my application: why is
this happening ? Some of the core web2py functions require to cache
on disk ?
