When a huge data file is created in a session, Tritcask fails to load it in the next sesssion.
This is caused by the use of mmap on a file bigger than the address space, in this case 9GB.
Here is the traceback:
File "<stdin>", line 1, in <module>
File "/usr/lib/python2.7/dist-packages/ubuntuone-client/ubuntuone/syncdaemon/tritcask.py", line 636, in __init__
self._build_keydir()
File "/usr/lib/python2.7/dist-packages/ubuntuone-client/ubuntuone/syncdaemon/tritcask.py", line 791, in _build_keydir
self._load_from_data(self.live_file)
File "/usr/lib/python2.7/dist-packages/ubuntuone-client/ubuntuone/syncdaemon/tritcask.py", line 819, in _load_from_data
for entry in data_file.iter_entries():
File "/usr/lib/python2.7/dist-packages/ubuntuone-client/ubuntuone/syncdaemon/tritcask.py", line 254, in iter_entries
fmmap = mmap.mmap(self.fd.fileno(), 0, access=mmap.ACCESS_READ)
ValueError: mmap length is too large
luciotorre
commented
10 months ago
When a huge data file is created in a session, Tritcask fails to load it in the next sesssion.
This is caused by the use of mmap on a file bigger than the address space, in this case 9GB.
Here is the traceback: