Bug #10445

[SDKs] Fix memory leak in Python SDK Collection class

Added by Tom Clegg over 2 years ago. Updated over 1 year ago.

Status:
New
Priority:
Normal
Assigned To:
-
Category:
SDKs
Target version:
Start date:
11/02/2016
Due date:
% Done:

0%

Estimated time:
Story points:
-

Description

Currently, each new CollectionReader creates its own API client, and Keep client, and block cache unless the caller supplies an API client object (and Keep client object?). If a caller creates 10 CollectionReaders and reads 64 MiB from each one, the program will use 640 MiB. Possibly due to HTTP KeepAlive behavior, Python does not reclaim memory even if the caller unreferences the CollectionReaders. For example, this script leaks memory and network connections:

import arvados

uuid = '......'
for i in range(20):
    cr = arvados.collection.CollectionReader(uuid)
    for fn in cr:
        f = cr.open(fn)
        f.read()
        f.close()

Proposed improvement:

Share block caches between auto-instantiated API clients that use the same settings.

leaktest.py (1.11 KB) leaktest.py Tom Clegg, 11/02/2016 03:32 PM

Related issues

Related to Arvados - Feature #10493: [Python SDK] Provide more control over pre-fetch behaviorNew11/09/2016

History

#1 Updated by Tom Clegg over 2 years ago

#2 Updated by Peter Amstutz over 2 years ago

The read cache is actually in KeepClient. Reads go through the block manager, but sharing the read cache is a matter of using a shared keepclient object.

(Technically, you could even create multiple KeepClient objects and initialize them with the same KeepBlockCache).

#3 Updated by Tom Clegg over 2 years ago

  • Description updated (diff)

#4 Updated by Tom Morris over 1 year ago

  • Target version set to Arvados Future Sprints

Also available in: Atom PDF