Added MySQL/Pacemaker troubleshooting faq section

This commit is contained in:
Pavel Lechenko 2013-07-31 23:54:47 +04:00
parent ded1624b6b
commit 6ff5fd40ee
3 changed files with 24 additions and 1 deletions

View File

@ -13,5 +13,6 @@ FAQ (Frequently Asked Questions)
.. include:: /pages/frequently-asked-questions/0070-common-technical-issues.rst
.. include:: /pages/frequently-asked-questions/0010-rabbitmq.rst
.. include:: /pages/frequently-asked-questions/0020-galera.rst
.. include:: /pages/frequently-asked-questions/0030-mysql-pacemaker.rst
.. include:: /pages/frequently-asked-questions/0080-other-questions.rst

View File

@ -0,0 +1,22 @@
Corosync crashes without network
--------------------------------
TOTEM, the network protocol used by Corosync exceeds its timeout. Corosync
sometimes crashes when networking is unavailable for a length of time.
Additionally, MySQL has stopped working.
**Workaround:**
#. Verify that corosync is really broken ``service corosync status``.
* You should see next error: ``corosync dead but pid file exists``
#. Start corosync manually ``service corosync start``.
#. Run ``ps -ef | grep mysql`` and kill ALL(!) **mysqld** and **mysqld_safe** processes.
#. Wait while pacemaker starts mysql processes again.
* You can check it with ``ps -ef | grep mysql`` command.
* If it doesn't start, run ``crm resource p_mysql`` start.
#. Check with ``crm status`` command that this host is part of the cluster and p_mysql is not within "Failed actions".

View File

@ -20,7 +20,7 @@ common:
pool_start: 10.49.54.225
pool_end: 10.49.54.239
segment_range: 900:999
tenant_network_type: vlan
tenant_network_type: gre
network_manager: nova.network.manager.FlatDHCPManager
auto_assign_floating_ip: true
quantum_netnode_on_cnt: true