how to execute kola tests

[Marshall Onsrud]

Question about how to properly execute kola tests. I've been trying for the past day or so to get any kola framework test to run on qemu. Varied attempts always result in ssh unreachable to the qemu under test. I can see a qemu process but am never able to connect to it via IP. However, I can boot the qemu image by hand and log into it so I know it is a valid image. Can anyone point me in the right direction?

sudo kola run  --verbose --platform qemu --qemu-image /mnt/host/source/src/build/images/amd64-usr/developer-1325.0.0+2017-03-01-1046-a1/coreos_developer_qemu_image.img

tester@localhost ~/trunk/chroot/usr/bin $ sudo kola run --verbose --platform qemu --qemu-image /mnt/host/source/src/build/images/amd64-usr/developer-1325.0.0+2017-03-01-1046-a1/coreos_developer_qemu_image.img

2017-03-01T16:28:51Z cli: Started logging at level INFO

2017-03-01T16:28:51Z util: dnsmasq[17132]: started, version 2.76 cachesize 150

2017-03-01T16:28:51Z util: dnsmasq[17132]: compile time options: IPv6 GNU-getopt no-DBus i18n IDN DHCP DHCPv6 no-scripts no-TFTP no-conntrack ipset no-auth no-DNSSEC loop-detect inotify

2017-03-01T16:28:51Z util: dnsmasq[17132]: warning: no upstream servers configured

2017-03-01T16:28:51Z util: dnsmasq-dhcp[17132]: DHCP, static leases only on 10.2.0.1, lease time 1h

2017-03-01T16:28:51Z util: dnsmasq-dhcp[17132]: DHCP, static leases only on 10.1.0.1, lease time 1h

2017-03-01T16:28:51Z util: dnsmasq-dhcp[17132]: DHCP, static leases only on 10.0.0.1, lease time 1h

2017-03-01T16:28:51Z util: dnsmasq-dhcp[17132]: DHCPv4-derived IPv6 names on fd02::

2017-03-01T16:28:51Z util: dnsmasq-dhcp[17132]: router advertisement on fd02::

2017-03-01T16:28:51Z util: dnsmasq-dhcp[17132]: DHCPv4-derived IPv6 names on fd01::

2017-03-01T16:28:51Z util: dnsmasq-dhcp[17132]: router advertisement on fd01::

2017-03-01T16:28:51Z util: dnsmasq-dhcp[17132]: DHCPv4-derived IPv6 names on fd00::

2017-03-01T16:28:51Z util: dnsmasq-dhcp[17132]: router advertisement on fd00::

2017-03-01T16:28:51Z util: dnsmasq-dhcp[17132]: IPv6 router advertisement enabled

2017-03-01T16:28:51Z util: dnsmasq[17132]: cleared cache

2017-03-01T16:28:51Z etcdserver: name = simple

2017-03-01T16:28:51Z etcdserver: data dir = /tmp/simple-etcd-624931551

2017-03-01T16:28:51Z etcdserver: member dir = /tmp/simple-etcd-624931551/member

2017-03-01T16:28:51Z etcdserver: heartbeat = 100ms

2017-03-01T16:28:51Z etcdserver: election = 1000ms

2017-03-01T16:28:51Z etcdserver: snapshot count = 0

2017-03-01T16:28:51Z etcdserver: advertise client URLs = http://10.0.0.1:33882,http://10.1.0.1:33882,http://10.2.0.1:33882,http://[fd00::1]:33882,http://[fd01::1]:33882,http://[fd02::1]:33882

2017-03-01T16:28:51Z etcdserver: initial advertise peer URLs = http://localhost:0

2017-03-01T16:28:51Z etcdserver: initial cluster = simple=http://localhost:0

2017-03-01T16:28:51Z etcdserver: starting member f3620cbbbe7cd162 in cluster 3bd8c6452814d712

2017-03-01T16:28:51Z raft: f3620cbbbe7cd162 became follower at term 0

2017-03-01T16:28:51Z raft: newRaft f3620cbbbe7cd162 [peers: [], term: 0, commit: 0, applied: 0, lastindex: 0, lastterm: 0]

2017-03-01T16:28:51Z raft: f3620cbbbe7cd162 became follower at term 1

2017-03-01T16:28:51Z etcdserver: set snapshot count to default 10000

2017-03-01T16:28:51Z etcdserver: starting server... [version: 3.0.3, cluster version: to_be_decided]

2017-03-01T16:28:51Z membership: added member f3620cbbbe7cd162 [http://localhost:0] to cluster 3bd8c6452814d712

2017-03-01T16:28:51Z network/ntp: Started NTP server on [::]:123

2017-03-01T16:28:52Z raft: f3620cbbbe7cd162 is starting a new election at term 1

2017-03-01T16:28:52Z raft: f3620cbbbe7cd162 became candidate at term 2

2017-03-01T16:28:52Z raft: f3620cbbbe7cd162 received vote from f3620cbbbe7cd162 at term 2

2017-03-01T16:28:52Z raft: f3620cbbbe7cd162 became leader at term 2

2017-03-01T16:28:52Z raft: raft.node: f3620cbbbe7cd162 elected leader f3620cbbbe7cd162 at term 2

2017-03-01T16:28:52Z etcdserver: published {Name:simple ClientURLs:[http://10.0.0.1:33882 http://10.1.0.1:33882 http://10.2.0.1:33882 http://[fd00::1]:33882 http://[fd01::1]:33882 http://[fd02::1]:33882]} to cluster 3bd8c6452814d712

2017-03-01T16:28:52Z etcdserver: setting up the initial cluster version to 3.0

2017-03-01T16:28:52Z membership: set the initial cluster version to 3.0

2017-03-01T16:28:52Z api: enabled capabilities for version 3.0

2017-03-01T16:33:20Z kola: creating new machine for semver check: ssh unreachable: dial tcp 10.0.0.2:22: getsockopt: no route to host

[Marshall Onsrud]

Maybe it is a problem with taking arguments like --qemu-image. in my build/images directory, I created a sym-link

called  coreos_production_image.bin and pointed it at my coreos_developer_image.bin . Afterward, I could run kola and see tests exectuing:

 sudo kola run --debug --verbose

If anyone knows how the cli params are supposed to work to pass a qemu image to be tested, keep me posted.

Thanks.

[Michael Marineau]

kola uses the basic prod image created by ./build_image prod; after that sudo kola run should work.

[Michael Marineau]

Also, as a quick FYI on kola I've just updated it in the SDK. The new
version includes a number of changes like always generating a
_kola_temp directory that includes data like the journals of every
machine that is launched to make troubleshooting issues more
practical.

[Marshall Onsrud]

This did work for me to build a prod image and run the tests against that. Thanks! All passed.