memory allocation errors on centos+source build

[Steven Dake]

From the build logs:

INFO:__main__:nova-network:Loaded plugins: fastestmirror, ovl, priorities, versionlock
INFO:__main__:nova-network:
INFO:__main__:nova-network:[Errno 12] Cannot allocate memory: '/var/cache/yum/x86_64/7/base/mirrorlist.txt'
INFO:__main__:nova-network:
INFO:__main__:nova-network:Removing intermediate container 0a98574de120
ERROR:__main__:nova-network:Error'd with the following message
ERROR:__main__:nova-network:The command '/bin/sh -c yum -y install initscripts && yum clean all' returned a non-zero code: 1
INFO:__main__:nova-network:Building

Disk is 37% full. Suspect it has to do with the memory limits set in the systemd file.

[Steven Dake (sdake)]

Another example:

INFO:__main__:rabbitmq:Step 3 : RUN yum -y install hostname https://github.com/rabbitmq/rabbitmq-server/releases/download/rabbitmq_v3_5_5/rabbitmq-server-3.5.5-3.noarch.rpm && yum clean all && rm -rf /var/lib/rabbitmq/* && curl -o /usr/lib/rabbitmq/lib/rabbitmq_server-3.5.5/plugins/rabbitmq_clusterer-3.5.x-189b3a81.ez http://www.rabbitmq.com/community-plugins/v3.5.x/rabbitmq_clusterer-3.5.x-189b3a81.ez
ERROR:__main__:rabbitmq:Error'd with the following message
ERROR:__main__:rabbitmq:mkdir /var/lib/docker/overlay/cf6a7b60dedef34f1d67d37b9fe86b58c1fbeb9fffa5cdd616e2bb648662e1f0-init/merged/dev/pts: cannot allocate memory
INFO:__main__:tempest:Building
INFO:__main__:keepalived:Loaded plugins: fastestmirror, ovl, priorities, versionlock
INFO:__main__:keepalived:Cannot open logfile /var/log/yum.log
INFO:__main__:keepalived:
INFO:__main__:keepalived:ovl: Error while doing RPMdb copy-up:
INFO:__main__:keepalived:[Errno 12] Cannot allocate memory: '/var/lib/rpm/.dbenv.lock'
INFO:__main__:keepalived:
INFO:__main__:keepalived:Could not create lock at /var/run/yum.pid: [Errno 12] Cannot allocate memory: '/var/run/yum.pid'
INFO:__main__:keepalived:
INFO:__main__:keepalived:
INFO:__main__:keepalived:Can't create lock file; exiting

[Steven Dake (sdake)]

Not sure what was causing this - possibly a network connectivity problem but it appears gone.