Repeatable locally and in latest Tempest full gate run:
21:39:35 ======================================================================
21:39:35 ERROR: Return an empty list for servers detail when no active servers
21:39:35 ----------------------------------------------------------------------
21:39:35 Traceback (most recent call last):
21:39:35 File "/opt/stack/tempest/tempest/tests/compute/test_list_servers_negative.py", line 331, in test_list_servers_detail_when_no_servers_running
21:39:35 self.client.wait_for_server_termination(server['id'])
21:39:35 File "/opt/stack/tempest/tempest/services/nova/json/servers_client.py", line 167, in wait_for_server_termination
21:39:35 raise exceptions.BuildErrorException(server_id=server_id)
21:39:35 BuildErrorException: Server b04debc5-943f-49a3-93df-a1e446bfb436 failed to build and is in ERROR status
21:39:35
21:39:35 ======================================================================
21:39:35 ERROR: Return an empty list when there are no active servers
21:39:35 ----------------------------------------------------------------------
21:39:35 Traceback (most recent call last):
21:39:35 File "/opt/stack/tempest/tempest/tests/compute/test_list_servers_negative.py", line 89, in test_list_servers_when_no_servers_running
21:39:35 self.client.wait_for_server_termination(server['id'])
21:39:35 File "/opt/stack/tempest/tempest/services/nova/json/servers_client.py", line 167, in wait_for_server_termination
21:39:35 raise exceptions.BuildErrorException(server_id=server_id)
21:39:35 BuildErrorException: Server b04debc5-943f-49a3-93df-a1e446bfb436 failed to build and is in ERROR status
Looking into the log files I found the following in n-sch's log:
Attempting to build 1 instance(s) from (pid=22058) schedule_run_instance /opt/stack/nova/nova/scheduler/filter_scheduler.py:66
2012-10-02 21:31:38 DEBUG nova.scheduler.filters.retry_filter [req-3010657d-1acc-483a-95c0-f9a377af1757 ListServersNegativeTest-user ListServersNegativeTest-tenant] Previously tried hosts: [u'devstack-1349212287']. (host=devstack-1349212287) from (pid=22058) host_passes /opt/stack/nova/nova/scheduler/filters/retry_filter.py:39
2012-10-02 21:31:38 DEBUG nova.scheduler.host_manager [req-3010657d-1acc-483a-95c0-f9a377af1757 ListServersNegativeTest-user ListServersNegativeTest-tenant] Host filter function <bound method RetryFilter.host_passes of <nova.scheduler.filters.retry_filter.RetryFilter object at 0x4b41e10>> failed for devstack-1349212287 from (pid=22058) passes_filters /opt/stack/nova/nova/scheduler/host_manager.py:253
2012-10-02 21:31:38 WARNING nova.scheduler.driver [req-3010657d-1acc-483a-95c0-f9a377af1757 ListServersNegativeTest-user ListServersNegativeTest-tenant] [instance: b04debc5-943f-49a3-93df-a1e446bfb436] Setting instance to ERROR state.
I'm not very familiar with this RetryFilter and when it was added to Nova. I will chat with folks on IRC to try and figure out why this is suddenly failing...
Relevant IRC conversation:
jaypipes> comstud, sandywalsh: around? I need some assistance trying to figure out the root cause of https:/ /bugs.launchpad .net/tempest/ +bug/1061167 servers_ negative failing two tests" [Critical, Confirmed] resource_ tracker. y) /bugs.launchpad .net/nova/ +bug/1061166 paste.openstack .org/show/ 21279/ 64.125. 181.92) has joined #openstack-dev manager [req-3010657d- 1acc-483a- 95c0-f9a377af17 57 ListServersNega tiveTest- user ListServersNega tiveTest- tenant] [instance: b04debc5- 943f-49a3- 93df-a1e446bfb4 36] Starting instance... fedora/ aeperezt) has joined #o...
<uvirtbot> Launchpad bug 1061167 in tempest "test_list_
<comstud> jaypipes: yes
<jaypipes> comstud, sandywalsh: has to do with "RetryFilter" in scheduler... I'm not familiar with that. Is it recently added?
<comstud> a couple months ago
<comstud> belliott added it
<comstud> but i'm familiar with it
<jaypipes> comstud: can you give me the sknny?
<comstud> essentially almost all instances -> ERROR are kicked back to the scheduler and retried
<jaypipes> I see.
<comstud> any exception caught in compute manager while building
<comstud> it's configurable... you can turn it off
<comstud> however.. it's needed to handle race conditions
<jaypipes> comstud: k.
<comstud> where we think compute has enough room
<comstud> but it really doesn't
<comstud> this goes along with the new resource tracking stuff
<comstud> (nova/compute/
<comstud> +p
* alunduil_ has quit (Ping timeout: 260 seconds)
* alunduil has quit (Ping timeout: 245 seconds)
<jaypipes> k
<rmk> Folsom horizon still has the admin warning banner across it. I thought all the bugs around this got sorted?
* dani4571 has quit (Quit: Leaving)
<comstud> jaypipes: There needs to be some better logging, etc, in the scheduler... and questionably a fault created...
<adam_g> is this a new one or is it already being tracked elsewhere? https:/
<uvirtbot> Launchpad bug 1061166 in nova "ec2 instance IDs are broken after folsom upgrade" [Undecided,New]
<comstud> so we know better which hosts were tried
<comstud> jaypipes: i think right now, you have to scan through your hosts for the instance uuid to find the traceback
<comstud> for the ultimate reason for failure
<jaypipes> comstud: that is what I have been doing :)
<comstud> yeah :)
<jaypipes> comstud: http://
<jaypipes> comstud: that
<jaypipes> 's the "error" that occurred on initial scheduling request
<jaypipes> comstud: which doens't make sense to me.
<comstud> weird
* danwent (~danwent@
<comstud> tempest is okay with us for xen
<comstud> at least on code within the past week
<jaypipes> comstud: and the line directly after that is even weirder...
<jaypipes> 2012-10-02 21:31:25 AUDIT nova.compute.
<comstud> lol
<jaypipes> comstud: so perhaps the QEMU error is indeed expected.
<comstud> same request id
<comstud> yeah
<jaypipes> comstud: and it's just output when an instance with the instance ID does not exist (checked prior to launch)
<comstud> what's the last log message for req-3010657d?
<comstud> or last log messages
<comstud> if something bombed out, i'd expect it to be a traceback
<comstud> they should still be logged before compute kicks it back to scheduler for retrying
* aeperezt (~aeperezt@