Open dgttm opened 5 years ago
UPDATE: as a workaround, once the rewrite engine is disabled (fast_forward=1) the issue is gone.
In the docker logs I found the a duplicate index warning. could this maybe lead to the described issue?
+++ nova-manage db sync
/var/lib/kolla/venv/local/lib/python2.7/site-packages/psycopg2/__init__.py:144: UserWarning: The psycopg2 wheel package will be renamed from release 2.8; in order to keep installing from binary please use "pip install psycopg2-binary" instead. For details see: <http://initd.org/psycopg/docs/install.html#binary-install-from-pypi>.
""")
/var/lib/kolla/venv/local/lib/python2.7/site-packages/pymysql/cursors.py:170: Warning: (1831, u"Duplicate index 'block_device_mapping_instance_uuid_virtual_name_device_name_idx' defined on the table 'nova_cell0.block_device_mapping'. This is deprecated and will be disallowed in a future release.")
result = self._query(query)
/var/lib/kolla/venv/local/lib/python2.7/site-packages/pymysql/cursors.py:170: Warning: (1831, u"Duplicate index 'uniq_instances0uuid' defined on the table 'nova_cell0.instances'. This is deprecated and will be disallowed in a future release.")
result = self._query(query)
Encountering same issue.
This is a reproducible issue.
[ ] A clear description of your issue:
Current Setup:
3-Nodes Cluster Setup for an Openstack Installation with external Percona XtraDB Galera DB and ProxySQL
Connection flow: Users -> 3-Node-ProxySQL Cluster with read/write split -> 3-Node-Percona XTRA DB Cluster
We encounter the issue, when deploying openstack with kolla-ansible, the databases are set up via ansible, so its a 100% reproducible setup.
While deploying openstack (e.g. bootstrapping keystone), there are DB transactions executed which fail with an error like:
I don't know exactly what the database actions are, but what I see is the following:
For us, it looks like there is kind of a race condition when doing the database transactions (something is read before its written or committed.. I'm not a DB admin, so sorry for not beeing more accurate).
we even disabled multiplexing, but still the same behaviour
[ ] The version of OS and ProxySQL
Description: Ubuntu 18.04.1 LTS
[ ] Every step to reproduce the issue
results in:
[ ] The error log
I only found the following errors in the proxysql.log (with one deployment)
Error Log from kolla-ansible when doing its bootstrapping
Thank you very much!!