[Starlingx-discuss] Question about patch upstreaming for e5712fe
Hi Matt, We're looking into the patch e5712fe, which indicates a possible inconsistency/mismatch of tag values between dnsmasq option files and dnsmasq command line.
From the commit message: "This seems to be happening because of how dhcp_rpc.py::_group_by_network_id would behave if the list of subnets coming from the database was ordered different or contained different values." After checking, we found that the latest Neutron seems to already have subnets sorted: https://github.com/openstack/neutron/blob/master/neutron/api/rpc/handlers/dh....
Would you please kindly help double check on this? Let me know if any question. Thank you! BR, Kailun
Hello Kailun, I agree that the change looks like it should address the same issue. The change from e5712fe may have been carried forward between rebases without realizing that it was addressed on the server side. Have you retested the scenario without the change for e5712fe? If things are working as expected, than it can be dropped. Regards, Matt From: "Qin, Kailun" <kailun.qin@intel.com> Date: Friday, October 12, 2018 at 6:33 AM To: "Peters, Matt" <Matt.Peters@windriver.com> Cc: "Le, Huifeng" <huifeng.le@intel.com>, "starlingx-discuss@lists.starlingx.io" <starlingx-discuss@lists.starlingx.io> Subject: Question about patch upstreaming for e5712fe Hi Matt, We’re looking into the patch e5712fe, which indicates a possible inconsistency/mismatch of tag values between dnsmasq option files and dnsmasq command line. From the commit message: “This seems to be happening because of how dhcp_rpc.py::_group_by_network_id would behave if the list of subnets coming from the database was ordered different or contained different values.” After checking, we found that the latest Neutron seems to already have subnets sorted: https://github.com/openstack/neutron/blob/master/neutron/api/rpc/handlers/dh.... Would you please kindly help double check on this? Let me know if any question. Thank you! BR, Kailun
Yes, your analysis is correct. Our patch (e5712fe) can be dropped because the issue has now been addressed by upstream commit (e45add7b07e5c72c43047d18e10af1c4ae307b0f). Regards, Allain Allain Legacy, Software Developer, Wind River direct 613.270.2279 fax 613.492.7870 skype allain.legacy 350 Terry Fox Drive, Suite 200, Ottawa, Ontario, K2K 2W5 [WIND]<http://www.windriver.com/> From: Qin, Kailun [mailto:kailun.qin@intel.com] Sent: Friday, October 12, 2018 6:33 AM To: Peters, Matt Cc: starlingx-discuss@lists.starlingx.io Subject: [Starlingx-discuss] Question about patch upstreaming for e5712fe Hi Matt, We're looking into the patch e5712fe, which indicates a possible inconsistency/mismatch of tag values between dnsmasq option files and dnsmasq command line.
From the commit message: "This seems to be happening because of how dhcp_rpc.py::_group_by_network_id would behave if the list of subnets coming from the database was ordered different or contained different values." After checking, we found that the latest Neutron seems to already have subnets sorted: https://github.com/openstack/neutron/blob/master/neutron/api/rpc/handlers/dh....
Would you please kindly help double check on this? Let me know if any question. Thank you! BR, Kailun
Hi Allain and Matt, Thanks for confirming. We've retested the scenario w/o the change e5712fe and cannot reproduce the issue. I think it's OK to drop the original patch e5712fe. BR, Kailun From: Legacy, Allain [mailto:Allain.Legacy@windriver.com] Sent: Friday, October 12, 2018 9:11 PM To: Qin, Kailun <kailun.qin@intel.com>; Peters, Matt <Matt.Peters@windriver.com> Cc: starlingx-discuss@lists.starlingx.io Subject: RE: Question about patch upstreaming for e5712fe Yes, your analysis is correct. Our patch (e5712fe) can be dropped because the issue has now been addressed by upstream commit (e45add7b07e5c72c43047d18e10af1c4ae307b0f). Regards, Allain Allain Legacy, Software Developer, Wind River direct 613.270.2279 fax 613.492.7870 skype allain.legacy 350 Terry Fox Drive, Suite 200, Ottawa, Ontario, K2K 2W5 [WIND]<http://www.windriver.com/> From: Qin, Kailun [mailto:kailun.qin@intel.com] Sent: Friday, October 12, 2018 6:33 AM To: Peters, Matt Cc: starlingx-discuss@lists.starlingx.io<mailto:starlingx-discuss@lists.starlingx.io> Subject: [Starlingx-discuss] Question about patch upstreaming for e5712fe Hi Matt, We're looking into the patch e5712fe, which indicates a possible inconsistency/mismatch of tag values between dnsmasq option files and dnsmasq command line.
From the commit message: "This seems to be happening because of how dhcp_rpc.py::_group_by_network_id would behave if the list of subnets coming from the database was ordered different or contained different values." After checking, we found that the latest Neutron seems to already have subnets sorted: https://github.com/openstack/neutron/blob/master/neutron/api/rpc/handlers/dh....
Would you please kindly help double check on this? Let me know if any question. Thank you! BR, Kailun
participants (3)
-
Legacy, Allain
-
Peters, Matt
-
Qin, Kailun