memory leak (during check --repair --add-lease)

[killyourtv]

While repairing a list of files I've had a problem with the memory usage for the tahoe process growing wildly and leading to OOM.

Attached are incident reports and screencaps of the webui's status page. I don't know if the statuses will tell you what the problem is but it'll be obvious that there is a problem.

Edit:

Oops, I neglected to report the versions:

allmydata-tahoe: 1.9.2,
foolscap: 0.6.4,
pycryptopp: 0.6.0.1206569328141510525648634803928199668821045408958,
zfec: 1.4.24,
Twisted: 12.0.0,
Nevow: 0.10.0,
zope.interface: unknown,
python: 2.7.3,
platform: Linux-debian_7.0-x86_64-64bit_ELF,
pyOpenSSL: 0.13,
simplejson: 2.5.2,
pycrypto: 2.6,
pyasn1: unknown,
mock: 0.8.0,
sqlite3: 2.6.0 [sqlite 3.7.13],
setuptools: 0.6 [distribute]

[killyourtv]

incident #1

[killyourtv]

incident 2

[killyourtv]

mutable file retrieve status

[daira]

Possibly a duplicate of #1824.

[kpreid]

Dumped the flogs; I see no similarities to #1824.