Comment 11 for bug 1493303

Revision history for this message
Tristan Cacqueray (tristan-cacqueray) wrote : Re: Swift proxy memory leak on unfinished read

I was able to reproduce this on saio instance of 8Go of ram, deployed with openstack-swift-2.3.0-2.fc23.noarch (which doesn't have 12d8a53). However it didn't cause a noticeable effect since the process was killed and automatically restarted without causing disruption to other services. Also the leak was quickly limited by the maximum number of open file and it took about 42k connections until the process reserved all memory.

Anyway this seems like a duplicate of bug 1466549 which is already public, can we mark this bug as duplicate and open it ?