I was able to get the hosted engine started manually via Virsh after re-creating a missing symlink in /var/run/vdsm/storage -- I later shut it down and am still having the same problem with ha broker starting. It appears that the problem *might* be with a corrupt ha metadata file, although gluster is not stating there is split brain on the engine volume

I'm seeing this:

ls -al /rhev/data-center/mnt/glusterSD/orchard0\:_engine/d70b171e-7488-4d52-8cad-bbc581dbf16e/ha_agent/
ls: cannot access /rhev/data-center/mnt/glusterSD/orchard0:_engine/d70b171e-7488-4d52-8cad-bbc581dbf16e/ha_agent/hosted-engine.metadata: Input/output error
total 0
drwxr-xr-x. 2 vdsm kvm 67 Dec 13 00:30 .
drwxr-xr-x. 6 vdsm kvm 64 Aug 6 2018 ..
lrwxrwxrwx. 1 vdsm kvm 132 Dec 13 00:30 hosted-engine.lockspace -> /var/run/vdsm/storage/d70b171e-7488-4d52-8cad-bbc581dbf16e/03a8ee8e-91f5-4e06-904b-9ed92a9706eb/db2699ce-6349-4020-b52d-8ab11d01e26d
l?????????? ? ? ? ? ? hosted-engine.metadata

Clearly showing some sort of issue with hosted-engine.metadata on the client mount.

on each node in /gluster_bricks I see this:

# ls -al /gluster_bricks/engine/engine/d70b171e-7488-4d52-8cad-bbc581dbf16e/ha_agent/
total 0
drwxr-xr-x. 2 vdsm kvm 67 Dec 13 00:31 .
drwxr-xr-x. 6 vdsm kvm 64 Aug 6 2018 ..
lrwxrwxrwx. 2 vdsm kvm 132 Dec 13 00:31 hosted-engine.lockspace -> /var/run/vdsm/storage/d70b171e-7488-4d52-8cad-bbc581dbf16e/03a8ee8e-91f5-4e06-904b-9ed92a9706eb/db2699ce-6349-4020-b52d-8ab11d01e26d
lrwxrwxrwx. 2 vdsm kvm 132 Dec 12 16:30 hosted-engine.metadata -> /var/run/vdsm/storage/d70b171e-7488-4d52-8cad-bbc581dbf16e/66bf05fa-bf50-45ec-98d8-d00002040317/a2250415-5ff0-42ab-8071-cd9d67c3048c

ls -al /var/run/vdsm/storage/d70b171e-7488-4d52-8cad-bbc581dbf16e/66bf05fa-bf50-45ec-98d8-d00002040317/a2250415-5ff0-42ab-8071-cd9d67c3048c
-rw-rw----. 1 vdsm kvm 1073741824 Dec 12 16:48 /var/run/vdsm/storage/d70b171e-7488-4d52-8cad-bbc581dbf16e/66bf05fa-bf50-45ec-98d8-d00002040317/a2250415-5ff0-42ab-8071-cd9d67c3048c

I'm not sure how to proceed at this point. Do I have data corruption, a gluster split-brain issue or something else? Maybe I just need to re-generate metadata for the hosted engine?

On Thu, Dec 12, 2019 at 6:36 PM Jayme <jaymef@gmail.com> wrote:

I'm running a three server HCI. Up and running on 4.3.7 with no problems. Today I updated to 4.3.8. Engine upgraded fine, rebooted. First host updated fine, rebooted and let all gluster volumes heal. Put second host in maintenance, upgraded successfully, rebooted. Waited for gluster volumes to heal for over an hour but the heal process was not completing. I tried restarting gluster services as well as the host with no success.

I'm in a state right now where there are pending heals on almost all of my volumes. Nothing is reporting split-brain, but the heals are not completing.

All vms are still currently running except hosted engine. Hosted engine was running but on the 2nd host I upgraded I was seeing errors such as:

Dec 12 16:34:39 orchard2 journal: ovirt-ha-agent ovirt_hosted_engine_ha.agent.hosted_engine.HostedEngine.config.vm ERROR Failed scanning for OVF_STORE due to Command Volume.getInfo with args {'storagepoolID': '00000000-0000-0000-0000-000000000000', 'storagedomainID': 'd70b171e-7488-4d52-8cad-bbc581dbf16e', 'volumeID': u'2632f423-ed89-43d9-93a9-36738420b866', 'imageID': u'd909dc74-5bbd-4e39-b9b5-755c167a6ee8'} failed:#012(code=201, message=Volume does not exist: (u'2632f423-ed89-43d9-93a9-36738420b866',))

I shut down the engine VM and attempted a manual heal on the engine volume. I cannot start the engine on any host now. I get:

The hosted engine configuration has not been retrieved from shared storage. Please ensure that ovirt-ha-agent is running and the storage server is reachable.

I'm seeing ovirt-ha-agent crashing on all three nodes:

Dec 12 18:30:48 orchard0 python: detected unhandled Python exception in '/usr/share/ovirt-hosted-engine-ha/ovirt-ha-broker'
Dec 12 18:30:48 orchard0 abrt-server: Duplicate: core backtrace
Dec 12 18:30:48 orchard0 abrt-server: DUP_OF_DIR: /var/tmp/abrt/Python-2019-03-14-21:02:52-44318
Dec 12 18:30:48 orchard0 abrt-server: Deleting problem directory Python-2019-12-12-18:30:48-23193 (dup of Python-2019-03-14-21:02:52-44318)
Dec 12 18:30:49 orchard0 vdsm[6087]: ERROR failed to retrieve Hosted Engine HA score '[Errno 2] No such file or directory'Is the Hosted Engine setup finished?
Dec 12 18:30:49 orchard0 systemd: ovirt-ha-broker.service: main process exited, code=exited, status=1/FAILURE
Dec 12 18:30:49 orchard0 systemd: Unit ovirt-ha-broker.service entered failed state.
Dec 12 18:30:49 orchard0 systemd: ovirt-ha-broker.service failed.
Dec 12 18:30:49 orchard0 systemd: ovirt-ha-broker.service holdoff time over, scheduling restart.
Dec 12 18:30:49 orchard0 systemd: Cannot add dependency job for unit lvm2-lvmetad.socket, ignoring: Unit is masked.
Dec 12 18:30:49 orchard0 systemd: Stopped oVirt Hosted Engine High Availability Communications Broker.

Here is what gluster volume heal info on engine looks like, it's similar on other volumes as well (although more heals pending on some of those):

gluster volume heal engine info
Brick gluster0:/gluster_bricks/engine/engine
/d70b171e-7488-4d52-8cad-bbc581dbf16e/images/d909dc74-5bbd-4e39-b9b5-755c167a6ee8/2632f423-ed89-43d9-93a9-36738420b866.meta
/d70b171e-7488-4d52-8cad-bbc581dbf16e/images/053171e4-f782-42d7-9115-c602beb3c826/627b8f93-5373-48bb-bd20-a308a455e082.meta
/d70b171e-7488-4d52-8cad-bbc581dbf16e/master/tasks/a9b11e33-9b93-46a0-a36e-85063fd53ebe.backup
/d70b171e-7488-4d52-8cad-bbc581dbf16e/dom_md/ids
Status: Connected
Number of entries: 4

Brick gluster1:/gluster_bricks/engine/engine
/d70b171e-7488-4d52-8cad-bbc581dbf16e/images/d909dc74-5bbd-4e39-b9b5-755c167a6ee8/2632f423-ed89-43d9-93a9-36738420b866.meta
/d70b171e-7488-4d52-8cad-bbc581dbf16e/master/tasks/a9b11e33-9b93-46a0-a36e-85063fd53ebe.backup
/d70b171e-7488-4d52-8cad-bbc581dbf16e/images/053171e4-f782-42d7-9115-c602beb3c826/627b8f93-5373-48bb-bd20-a308a455e082.meta
/d70b171e-7488-4d52-8cad-bbc581dbf16e/dom_md/ids
Status: Connected
Number of entries: 4

Brick gluster2:/gluster_bricks/engine/engine
Status: Connected
Number of entries: 0

I don't see much in vdsm.log and gluster logs look fairly normal to me, I'm not seeing any obvious errors in the gluster logs.

As far as I can tell the underlying storage is fine. Why are my gluster volumes not healing and why is self-hosted engine failing to start?

More agent and broker logs:

==> agent.log <==
MainThread::ERROR::2019-12-12 18:36:09,056::hosted_engine::559::ovirt_hosted_engine_ha.agent.hosted_engine.HostedEngine::(_initialize_broker) Failed to start necessary monitors
MainThread::ERROR::2019-12-12 18:36:09,058::agent::144::ovirt_hosted_engine_ha.agent.agent.Agent::(_run_agent) Traceback (most recent call last):
File "/usr/lib/python2.7/site-packages/ovirt_hosted_engine_ha/agent/agent.py", line 131, in _run_agent
return action(he)
File "/usr/lib/python2.7/site-packages/ovirt_hosted_engine_ha/agent/agent.py", line 55, in action_proper
return he.start_monitoring()
File "/usr/lib/python2.7/site-packages/ovirt_hosted_engine_ha/agent/hosted_engine.py", line 432, in start_monitoring
self._initialize_broker()
File "/usr/lib/python2.7/site-packages/ovirt_hosted_engine_ha/agent/hosted_engine.py", line 556, in _initialize_broker
m.get('options', {}))
File "/usr/lib/python2.7/site-packages/ovirt_hosted_engine_ha/lib/brokerlink.py", line 89, in start_monitor
).format(t=type, o=options, e=e)
RequestError: brokerlink - failed to start monitor via ovirt-ha-broker: [Errno 2] No such file or directory, [monitor: 'network', options: {'tcp_t_address': None, 'network_test': None, 'tcp_t_port': None, 'addr': '10.11.0.254'}]

MainThread::ERROR::2019-12-12 18:36:09,058::agent::145::ovirt_hosted_engine_ha.agent.agent.Agent::(_run_agent) Trying to restart agent
MainThread::ERROR::2019-12-12 18:36:19,619::hosted_engine::559::ovirt_hosted_engine_ha.agent.hosted_engine.HostedEngine::(_initialize_broker) Failed to start necessary monitors
MainThread::ERROR::2019-12-12 18:36:19,619::agent::144::ovirt_hosted_engine_ha.agent.agent.Agent::(_run_agent) Traceback (most recent call last):
File "/usr/lib/python2.7/site-packages/ovirt_hosted_engine_ha/agent/agent.py", line 131, in _run_agent
return action(he)
File "/usr/lib/python2.7/site-packages/ovirt_hosted_engine_ha/agent/agent.py", line 55, in action_proper
return he.start_monitoring()
File "/usr/lib/python2.7/site-packages/ovirt_hosted_engine_ha/agent/hosted_engine.py", line 432, in start_monitoring
self._initialize_broker()
File "/usr/lib/python2.7/site-packages/ovirt_hosted_engine_ha/agent/hosted_engine.py", line 556, in _initialize_broker
m.get('options', {}))
File "/usr/lib/python2.7/site-packages/ovirt_hosted_engine_ha/lib/brokerlink.py", line 89, in start_monitor
).format(t=type, o=options, e=e)
RequestError: brokerlink - failed to start monitor via ovirt-ha-broker: [Errno 2] No such file or directory, [monitor: 'network', options: {'tcp_t_address': None, 'network_test': None, 'tcp_t_port': None, 'addr': '10.11.0.254'}]

MainThread::ERROR::2019-12-12 18:36:19,619::agent::145::ovirt_hosted_engine_ha.agent.agent.Agent::(_run_agent) Trying to restart agent
MainThread::ERROR::2019-12-12 18:36:30,568::hosted_engine::559::ovirt_hosted_engine_ha.agent.hosted_engine.HostedEngine::(_initialize_broker) Failed to start necessary monitors
MainThread::ERROR::2019-12-12 18:36:30,570::agent::144::ovirt_hosted_engine_ha.agent.agent.Agent::(_run_agent) Traceback (most recent call last):
File "/usr/lib/python2.7/site-packages/ovirt_hosted_engine_ha/agent/agent.py", line 131, in _run_agent
return action(he)
File "/usr/lib/python2.7/site-packages/ovirt_hosted_engine_ha/agent/agent.py", line 55, in action_proper
return he.start_monitoring()
File "/usr/lib/python2.7/site-packages/ovirt_hosted_engine_ha/agent/hosted_engine.py", line 432, in start_monitoring
self._initialize_broker()
File "/usr/lib/python2.7/site-packages/ovirt_hosted_engine_ha/agent/hosted_engine.py", line 556, in _initialize_broker
m.get('options', {}))
File "/usr/lib/python2.7/site-packages/ovirt_hosted_engine_ha/lib/brokerlink.py", line 89, in start_monitor
).format(t=type, o=options, e=e)
RequestError: brokerlink - failed to start monitor via ovirt-ha-broker: [Errno 2] No such file or directory, [monitor: 'network', options: {'tcp_t_address': None, 'network_test': None, 'tcp_t_port': None, 'addr': '10.11.0.254'}]

MainThread::ERROR::2019-12-12 18:36:30,570::agent::145::ovirt_hosted_engine_ha.agent.agent.Agent::(_run_agent) Trying to restart agent
MainThread::ERROR::2019-12-12 18:36:41,581::hosted_engine::559::ovirt_hosted_engine_ha.agent.hosted_engine.HostedEngine::(_initialize_broker) Failed to start necessary monitors
MainThread::ERROR::2019-12-12 18:36:41,583::agent::144::ovirt_hosted_engine_ha.agent.agent.Agent::(_run_agent) Traceback (most recent call last):
File "/usr/lib/python2.7/site-packages/ovirt_hosted_engine_ha/agent/agent.py", line 131, in _run_agent
return action(he)
File "/usr/lib/python2.7/site-packages/ovirt_hosted_engine_ha/agent/agent.py", line 55, in action_proper
return he.start_monitoring()
File "/usr/lib/python2.7/site-packages/ovirt_hosted_engine_ha/agent/hosted_engine.py", line 432, in start_monitoring
self._initialize_broker()
File "/usr/lib/python2.7/site-packages/ovirt_hosted_engine_ha/agent/hosted_engine.py", line 556, in _initialize_broker
m.get('options', {}))
File "/usr/lib/python2.7/site-packages/ovirt_hosted_engine_ha/lib/brokerlink.py", line 89, in start_monitor
).format(t=type, o=options, e=e)
RequestError: brokerlink - failed to start monitor via ovirt-ha-broker: [Errno 2] No such file or directory, [monitor: 'network', options: {'tcp_t_address': None, 'network_test': None, 'tcp_t_port': None, 'addr': '10.11.0.254'}]

MainThread::ERROR::2019-12-12 18:36:41,583::agent::145::ovirt_hosted_engine_ha.agent.agent.Agent::(_run_agent) Trying to restart agent