EOF error during SSH connection

[Salvatore Orlando]

This has been observed during a tempest run: http://paste.openstack.org/show/63009/

It might be a paramiko issue, and might not happen anymore, but it sound like another intermittent failure that need to be tracked with an elastic recheck query.

Full logs: http://logs.openstack.org/90/67390/6/check/check-tempest-dsvm-neutron-isolated/552fb81

[Darragh O'Reilly (darragh-oreilly)]

I don't see the ssh problem in the logs - I think the ssh connection via the floatingip was ok and "[chan 1] EOF received (1)" is normal?

Is this not the problem http://logs.openstack.org/90/67390/6/check/check-tempest-dsvm-neutron-isolated/552fb81/console.html.gz#_2014-02-07_01_58_54_816

  Traceback (most recent call last):
   File "tempest/scenario/test_cross_tenant_connectivity.py", line 481, in test_cross_tenant_traffic
     self._verify_mac_addr(self.tenants[tenant_id])
   File "tempest/scenario/test_cross_tenant_connectivity.py", line 470, in _verify_mac_addr
     server_ip = self._get_server_ip(tenant.access_point)
   File "tempest/scenario/test_cross_tenant_connectivity.py", line 327, in _get_server_ip
     return server.networks[network_name][0]
   KeyError: u'network-smoke--1752694621'

It seems the "addresses" attribute was empty on the last GET. http://logs.openstack.org/90/67390/6/check/check-tempest-dsvm-neutron-isolated/552fb81/console.html.gz#_2014-02-07_01_58_54_692

RESP BODY: {"server": {"status": "ACTIVE", "updated": "2014-02-07T01:47:45Z", "hostId": "6a46aadcb9b335473c51951cf61a4a8215d59e4ea00e5b632d6276fc", "addresses": {} ...

I checked other logs for a run without this failure and "addresses" had a dict that included the fixed ip and mac.

[Darragh O'Reilly (darragh-oreilly)]

It has only happened 4 times in the last 7 days:
http://logstash.openstack.org/#eyJzZWFyY2giOiJtZXNzYWdlOiBcInNlcnZlci5uZXR3b3Jrc1tuZXR3b3JrX25hbWVdXCIiLCJmaWVsZHMiOltdLCJvZmZzZXQiOjAsInRpbWVmcmFtZSI6IjYwNDgwMCIsImdyYXBobW9kZSI6ImNvdW50IiwidGltZSI6eyJ1c2VyX2ludGVydmFsIjowfSwic3RhbXAiOjEzOTIwNDM2ODUyNTN9

There is a useful comment here:
https://github.com/openstack/tempest/blob/103aa1cab91456887bdb457e18c4b4270434ed10/tempest/scenario/manager.py#L472
# The instance retrieved on creation is missing network
# details, necessitating retrieval after it becomes active to
# ensure correct details.

It seems that sometimes does not work.

[Salvatore Orlando (salvatore-orlando)]

Apparently happens more often with parallel tests:
http://logstash.openstack.org/#eyJzZWFyY2giOiJtZXNzYWdlOidGaWxlIFwiL3Vzci9sb2NhbC9saWIvcHl0aG9uMi43L2Rpc3QtcGFja2FnZXMvcGFyYW1pa28vdHJhbnNwb3J0LnB5XCInIEFORCBtZXNzYWdlOlwic3RhcnRfY2xpZW50XCIgQU5EIGJ1aWxkX25hbWU6XCJjaGVjay10ZW1wZXN0LWRzdm0tbmV1dHJvbi1mdWxsXCIiLCJmaWVsZHMiOltdLCJvZmZzZXQiOjAsInRpbWVmcmFtZSI6IjE3MjgwMCIsImdyYXBobW9kZSI6ImNvdW50IiwidGltZSI6eyJ1c2VyX2ludGVydmFsIjowfSwic3RhbXAiOjEzOTMwOTgxNTY5MTcsIm1vZGUiOiIiLCJhbmFseXplX2ZpZWxkIjoiIn0=

NOTE - I'm using a different query as well

[Eugene Nikanorov (enikanorov)]

Is this bug still valid?

[Launchpad Janitor (janitor)]

[Expired for neutron because there has been no activity for 60 days.]