moby: Docker gets killed by SIGPIPE

Hi!

From time to time, docker running on my host is getting killed by SIGPIPE. I am running it through systemd:

docker.io.service - Docker Application Container Engine
   Loaded: loaded (/lib/systemd/system/docker.io.service; enabled)
   Active: inactive (dead) since jeu. 2014-07-17 19:33:03 CEST; 34s ago
     Docs: http://docs.docker.io
 Main PID: 29636 (code=killed, signal=PIPE)

juil. 16 00:03:34 neo docker[29636]: [466c6525] -job images() = OK (0)
juil. 16 00:03:34 neo docker[29636]: 2014/07/16 00:03:34 GET /v1.12/containers/json
juil. 16 00:03:34 neo docker[29636]: [466c6525] +job containers()
juil. 16 00:03:34 neo docker[29636]: [466c6525] -job containers() = OK (0)
juil. 16 08:48:50 neo docker[29636]: 2014/07/16 08:48:50 GET /v1.12/images/json
juil. 16 08:48:50 neo docker[29636]: [466c6525] +job images()
juil. 16 08:48:51 neo docker[29636]: [466c6525] -job images() = OK (0)
juil. 16 08:48:51 neo docker[29636]: 2014/07/16 08:48:51 GET /v1.12/containers/json
juil. 16 08:48:51 neo docker[29636]: [466c6525] +job containers()
juil. 16 08:48:51 neo docker[29636]: [466c6525] -job containers() = OK (0)

This is a very old problem that I had in early versions. I never investigated it but since it is still here, I am opening this bug report. Usually, no containers are running.

However, note that I am using 1.0.0 (from Debian unstable). 1.1.0 is not packaged yet. I have already reported the problem to Debian but got no answer yet.