Re: [openstack-dev] [sqlalchemy][db][oslo.db][mistral] Is there a recommended MySQL driver for OpenStack projects?

2018-07-03 Thread Renat Akhmerov
> > If you have a scaling issue that may be solved by eventlet, that's > one thing, but please don't adopt eventlet just because a lot of > other projects have. We've tried several times to minimize our > reliance on eventlet because new releases tend to introduce bugs. > > Have you tried the

Re: [openstack-dev] [sqlalchemy][db][oslo.db][mistral] Is there a recommended MySQL driver for OpenStack projects?

2018-07-03 Thread Jay Pipes
On 07/03/2018 08:47 AM, Doug Hellmann wrote: If you have a scaling issue that may be solved by eventlet, that's one thing, but please don't adopt eventlet just because a lot of other projects have. We've tried several times to minimize our reliance on eventlet because new releases tend to

Re: [openstack-dev] [sqlalchemy][db][oslo.db][mistral] Is there a recommended MySQL driver for OpenStack projects?

2018-07-03 Thread Doug Hellmann
Excerpts from Renat Akhmerov's message of 2018-07-03 18:33:44 +0700: > Hi, > > We’ve tried to address the bug [1] which is essentially caused > by the fact that we saw that MySQLDb driver wasn’t compatible with > eventlet’s green threads. In a nutshell, when we used the “eventlet” > RPC executor

Re: [openstack-dev] [sqlalchemy][db][oslo.db][mistral] Is there a recommended MySQL driver for OpenStack projects?

2018-07-03 Thread Jeremy Stanley
On 2018-07-03 18:33:44 +0700 (+0700), Renat Akhmerov wrote: [...] > So, I’m trying to understand if we have a “community recommended” > (or may be even mandatory) Mysql driver to use and what > consequences of the driver choice are. I’d appreciate any help > with clarifying this (may be links to

[openstack-dev] [sqlalchemy][db][oslo.db][mistral] Is there a recommended MySQL driver for OpenStack projects?

2018-07-03 Thread Renat Akhmerov
Hi, We’ve tried to address the bug [1] which is essentially caused by the fact that we saw that MySQLDb driver wasn’t compatible with eventlet’s green threads. In a nutshell, when we used the “eventlet” RPC executor (see [2]), the system would get stuck once in a while when dispatching green