[openstack-dev] [qa][python-bindings] Reries on HTTP error in python bindings
Hi all, I've been working on a patch to ensure that all tempest tests are able to handle rate limiting errors with automatic retries [0]. API tests already implement retries, however scenario tests don't, the main reason being that scenario tests rely on the official python bindings, most of which do not handle retries (with the exception of the swift client). My approach so far has been to monkey-patch python-bindings in tempest to add the retry logic, however a few pointed out that the right place to implement the retries would be in the python bindings directly. I filed a bug [1] about implementing retries in the various python bindings, and I'd like to get some feedback from the community. Even if we have enough consensus about [1], I expect it might be a while until all python bindings are updated. In the meanwhile I think we could go ahead with [0], and then start removing the monkey-patch as soon as retries are implemented on the client side. andrea [0] https://review.openstack.org/#/c/61978/ [1] https://bugs.launchpad.net/python-novaclient/+bug/1263069 -- Andrea Frittoli IaaS Systems Engineering Team HP Cloud Services smime.p7s Description: S/MIME cryptographic signature ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [qa][python-bindings] Reries on HTTP error in python bindings
Frittoli, I don't think that this is a good idea. Because the main role of tempest is to ensure that cloud works. So if it doesn't work we should change something in deployment configuration or code to fix it. Not just hide with monkey patches in tempest. Probably you should try to implement retry logic in python clients? Best regards, Boris Pavlovic On Wed, Jan 15, 2014 at 10:00 PM, Joe Gordon joe.gord...@gmail.com wrote: On Wed, Jan 15, 2014 at 5:28 AM, Frittoli, Andrea (Cloud Services) fritt...@hp.com wrote: Hi all, I’ve been working on a patch to ensure that all tempest tests are able to handle rate limiting errors with automatic retries [0]. API tests already implement retries, however scenario tests don’t, the main reason being that scenario tests rely on the official python bindings, most of which do not handle retries (with the exception of the swift client). My approach so far has been to monkey-patch python-bindings in tempest to add the retry logic, however a few pointed out that the right place to implement the retries would be in the python bindings directly. I filed a bug [1] about implementing retries in the various python bindings, and I’d like to get some feedback from the community. Even if we have enough consensus about [1], I expect it might be a while until all python bindings are updated. In the meanwhile I think we could go ahead with [0], and then start removing the monkey-patch as soon as retries are implemented on the client side. Besides monkey patching being scary and dangerous, monkey patching the clients in tempest means that we are now gating the clients on if the monkey patch is still valid. So we can't change the clients in a way that would break the monkey patch. Although we could put the monkey patching in a try/except clause as long as we don't exercise this in the gate. andrea [0] https://review.openstack.org/#/c/61978/ [1] https://bugs.launchpad.net/python-novaclient/+bug/1263069 -- Andrea Frittoli IaaS Systems Engineering Team HP Cloud Services ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [qa][python-bindings] Reries on HTTP error in python bindings
Thanks Joe, Good point, gating the clients with monkey-patched clients could become nasty. I'll pursue the changing the clients option only then. andrea From: Joe Gordon [mailto:joe.gord...@gmail.com] Sent: 15 January 2014 18:01 To: OpenStack Development Mailing List (not for usage questions) Subject: Re: [openstack-dev] [qa][python-bindings] Reries on HTTP error in python bindings On Wed, Jan 15, 2014 at 5:28 AM, Frittoli, Andrea (Cloud Services) fritt...@hp.com wrote: Hi all, I've been working on a patch to ensure that all tempest tests are able to handle rate limiting errors with automatic retries [0]. API tests already implement retries, however scenario tests don't, the main reason being that scenario tests rely on the official python bindings, most of which do not handle retries (with the exception of the swift client). My approach so far has been to monkey-patch python-bindings in tempest to add the retry logic, however a few pointed out that the right place to implement the retries would be in the python bindings directly. I filed a bug [1] about implementing retries in the various python bindings, and I'd like to get some feedback from the community. Even if we have enough consensus about [1], I expect it might be a while until all python bindings are updated. In the meanwhile I think we could go ahead with [0], and then start removing the monkey-patch as soon as retries are implemented on the client side. Besides monkey patching being scary and dangerous, monkey patching the clients in tempest means that we are now gating the clients on if the monkey patch is still valid. So we can't change the clients in a way that would break the monkey patch. Although we could put the monkey patching in a try/except clause as long as we don't exercise this in the gate. andrea [0] https://review.openstack.org/#/c/61978/ [1] https://bugs.launchpad.net/python-novaclient/+bug/1263069 -- Andrea Frittoli IaaS Systems Engineering Team HP Cloud Services ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev smime.p7s Description: S/MIME cryptographic signature ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
