CI: services memory fails

[Sagi (Sergey) Shnaidman]

There are a lot of puppet memory problems recently, mostly in HA jobs (but not limited to)

Building remotely on tripleo-centos-7-tripleo-test-cloud-rh1-3344848

HA:

Error: Could not prefetch keystone_domain provider 'openstack': Cannot allocate memory - fork(2)
http://logs.openstack.org/07/353307/2/check-tripleo/gate-tripleo-ci-centos-7-ovb-ha/dee6294/logs/postci.txt.gz

Error: Could not set 'present' on ensure: Cannot allocate memory - fork(2)
http://logs.openstack.org/64/352364/8/check-tripleo/gate-tripleo-ci-centos-7-ovb-ha/d6271f0/logs/postci.txt.gz

non-HA:

Error: /Stage[main]/Heat::Api_cloudwatch/Service[heat-api-cloudwatch]: Could not evaluate: Cannot allocate memory - fork(2)
http://logs.openstack.org/72/329872/15/check-tripleo/gate-tripleo-ci-centos-7-ovb-nonha/91d3244/logs/postci.txt.gz

+ free -h
              total used free shared buff/cache available
Mem: 5.8G 5.4G 169M 624K 309M 224M
Swap: 8.0G 62M 7.9G

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
  872 heat 20 0 690328 460072 3392 S 0.0 7.5 5:23.60 heat-engine
  873 heat 20 0 527468 297088 3416 S 0.0 4.9 6:24.99 heat-engine
  874 heat 20 0 521608 291536 3580 S 0.0 4.8 5:16.55 heat-engine
  875 heat 20 0 514928 284672 3396 S 0.0 4.7 4:19.51 heat-engine
25204 mysql 20 0 1650920 283484 6144 S 0.0 4.6 1:38.52 mysqld
 1006 nova 20 0 842320 193192 4716 S 0.0 3.2 1:42.27 httpd

[Sagi (Sergey) Shnaidman (sshnaidm)]

With adding memory to overcloud vms it passed
https://review.openstack.org/#/c/353655/

[Sagi (Sergey) Shnaidman (sshnaidm)]

All ovb-ha jobs are failing again with puppet memory problem.

[Sagi (Sergey) Shnaidman (sshnaidm)]

https://review.openstack.org/#/c/357532/

[Sagi (Sergey) Shnaidman (sshnaidm)]

After increasing memory for undercloud to 8GB it started passing again.