| opendevreview | zhangtongjian proposed openstack/neutron master: add has lock judgment for OvnNbSynchronizer https://review.opendev.org/c/openstack/neutron/+/800751 | 01:18 |
|---|---|---|
| opendevreview | Manu B proposed openstack/neutron-lib master: Introduce new bgp_associations API definition https://review.opendev.org/c/openstack/neutron-lib/+/792774 | 05:13 |
| *** iurygregory_ is now known as iurygregory | 06:38 | |
| zhangtongjian | ralonsoh, hello | 07:06 |
| ralonsoh | <zhangtongjian hi (are you still there?) | 07:24 |
| zhangtongjian | ralonsoh, yes | 07:27 |
| ralonsoh | so please, go on | 07:28 |
| zhangtongjian | to the patch https://review.opendev.org/c/openstack/neutron/+/800751, i have do some tests, the set lock is asynchronously, can not has lock right now | 07:29 |
| zhangtongjian | ralonson, so i think there is a problem with the writing of the lock immediately obtained after the lock is added | 07:30 |
| *** rpittau|afk is now known as rpittau | 07:31 | |
| zhangtongjian | ralonsoh, | 07:32 |
| zhangtongjian | Is it reasonable to do some sleep times? | 07:32 |
| ralonsoh | zhangtongjian, right, "is_lock_contended" will be set by the reply from the DB server | 07:34 |
| ralonsoh | let me check | 07:35 |
| zhangtongjian | ok | 07:35 |
| *** dmellado_ is now known as dmellado | 07:41 | |
| ralonsoh | zhangtongjian, what you have right now | 07:42 |
| ralonsoh | https://review.opendev.org/c/openstack/neutron/+/800751/7/neutron/plugins/ml2/drivers/ovn/mech_driver/mech_driver.py#335 | 07:42 |
| ralonsoh | should work | 07:42 |
| ralonsoh | right? | 07:42 |
| ralonsoh | has_lock is set when the lock is defined | 07:42 |
| ralonsoh | is_lock_contended will be false if you are the first server requesting defining and requesting this lock | 07:42 |
| zhangtongjian | yes, if i set lock , and then the has_lock is false, if i sleep 1s , the has_lcok is true | 07:45 |
| ralonsoh | zhangtongjian, so this is why I said in https://review.opendev.org/c/openstack/neutron/+/800751/6/neutron/plugins/ml2/drivers/ovn/mech_driver/mech_driver.py#336 | 07:47 |
| ralonsoh | that has_lock is not needed | 07:47 |
| ralonsoh | use only "is_lock_contended" | 07:47 |
| ralonsoh | if false, no other server is requesting the lock | 07:47 |
| ralonsoh | if true, there is another server accessing to the DB | 07:47 |
| zhangtongjian | yes | 07:48 |
| zhangtongjian | but if is_lock_contended is false, must the current IDL get the lock? | 07:49 |
| ralonsoh | it will, same as in DBInconsistenciesPeriodics.has_lock | 07:50 |
| ralonsoh | @property | 07:50 |
| ralonsoh | def has_lock(self): | 07:50 |
| ralonsoh | return not self._idl.is_lock_contended | 07:50 |
| zhangtongjian | so i think could this "DBInconsistenciesPeriodics.has_lock" be wrong? | 07:55 |
| ralonsoh | why? | 07:56 |
| zhangtongjian | no other server is requesting the lock, i am not sure if the current IDL must has get the lock | 07:58 |
| ralonsoh | if no one has requested the lock | 07:59 |
| ralonsoh | "is_lock_contended" will be false | 07:59 |
| zhangtongjian | yes | 07:59 |
| ralonsoh | when you check for "is_lock_contended" and other server is accessing to the DB, "is_lock_contended" will be True | 07:59 |
| ralonsoh | this is set directly by the DB server on the IDL instances | 07:59 |
| zhangtongjian | i have a question: what is the difference between the "has_lock" and "is_lock_contended"? | 08:01 |
| zhangtongjian | ralonsoh, are you still there? | 08:15 |
| ralonsoh | zhangtongjian, both parameters are set by the server | 08:15 |
| ralonsoh | zhangtongjian, has_lock will tell you if you have the lock | 08:16 |
| ralonsoh | "is_lock_contended" will tell you if you can't get the lock | 08:16 |
| ralonsoh | if "is_lock_contended" is False, then you can continue and access to the DB | 08:16 |
| zhangtongjian | so what method should use is right for me ? | 08:17 |
| ralonsoh | you should check only for "is_lock_contended" | 08:18 |
| ralonsoh | if not idl.is_lock_contended: ... | 08:19 |
| zhangtongjian | another a question: the set lock is asynchronously , and the "idl.is_lock_contended" can return false Immediately? | 08:22 |
| zhangtongjian | ralonsoh, maybe i do some sleep times? | 08:24 |
| ralonsoh | zhangtongjian, never user sleeps, in any case, active waits | 08:24 |
| ralonsoh | "idl.is_lock_contended" is False by default | 08:24 |
| ralonsoh | check OVS code | 08:25 |
| ralonsoh | https://github.com/openvswitch/ovs/blob/master/python/ovs/db/idl.py#L171 | 08:25 |
| zhangtongjian | ok, let me check | 08:27 |
| zhangtongjian | yes, default is false | 08:28 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: Remove ``ConfDriver`` code https://review.opendev.org/c/openstack/neutron/+/790999 | 08:50 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: Add a privsep context only for link commands https://review.opendev.org/c/openstack/neutron/+/800686 | 09:44 |
| opendevreview | Merged openstack/neutron-dynamic-routing stable/wallaby: Fix TypeError when formatting BGP IP address. https://review.opendev.org/c/openstack/neutron-dynamic-routing/+/786324 | 09:52 |
| opendevreview | Lajos Katona proposed openstack/networking-odl master: WIP: use payloads for PORT AFTER_UPDATE and SECURITY_GROUPS events https://review.opendev.org/c/openstack/networking-odl/+/800612 | 10:03 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: [OVN][Placement] Drive binding by placement allocation https://review.opendev.org/c/openstack/neutron/+/786478 | 10:17 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: Remove "_get_network_lock_id" compatibility method https://review.opendev.org/c/openstack/neutron/+/800967 | 10:58 |
| opendevreview | Lajos Katona proposed openstack/networking-odl master: WIP: use payloads for PORT AFTER_UPDATE and SECURITY_GROUPS events https://review.opendev.org/c/openstack/networking-odl/+/800612 | 11:24 |
| opendevreview | Chris MacNaughton proposed openstack/neutron stable/ussuri: Provide integer number to arping "-w" parameter https://review.opendev.org/c/openstack/neutron/+/800868 | 11:54 |
| opendevreview | Lajos Katona proposed openstack/networking-odl master: Use payloads for PORT SUBNET and SECURITY_GROUPS events https://review.opendev.org/c/openstack/networking-odl/+/800612 | 12:01 |
| icey | hey ralonsoh thanks for the quick code review on that backport proposal :) | 12:02 |
| ralonsoh | icey, yw | 12:02 |
| opendevreview | Merged openstack/networking-ovn stable/train: [OVN] Fix ML2/OVN + Neutron DHCP agent use case https://review.opendev.org/c/openstack/networking-ovn/+/800896 | 12:20 |
| opendevreview | Lucas Alvares Gomes proposed openstack/neutron master: [OVN] Make external ports aware of network AZs https://review.opendev.org/c/openstack/neutron/+/762550 | 13:08 |
| opendevreview | Lucas Alvares Gomes proposed openstack/neutron master: [OVN] Document Network Availability Zones https://review.opendev.org/c/openstack/neutron/+/788040 | 13:08 |
| pjakuszew | Just wanted to confirm, but this value is expressed in seconds, right? https://opendev.org/openstack/neutron-vpnaas/src/branch/stable/rocky/neutron_vpnaas/services/vpn/device_drivers/ipsec.py#L58-L60 | 13:20 |
| pjakuszew | s/but// | 13:27 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: Remove ``ConfDriver`` code https://review.opendev.org/c/openstack/neutron/+/790999 | 13:27 |
| ygk_12345 | Is anyone familiar with routed networks ? | 13:27 |
| ygk_12345 | I need some information regarding that | 13:27 |
| ygk_12345 | What is exactly multi-segment layer-3 network. In the context of routed networks ? | 13:28 |
| ygk_12345 | Is there such a thing as l3 layer segmention ? | 13:28 |
| lajoskatona | pjakuszew: Hi, after running through the code yes it's in seconds: https://opendev.org/openstack/oslo.service/src/branch/master/oslo_service/loopingcall.py#L118 | 13:32 |
| ralonsoh | ygk_12345, did you check https://docs.openstack.org/neutron/pike/admin/config-routed-networks.html? | 13:34 |
| ygk_12345 | I am looking to understand why we need segmentation at layer 3 when we can segment using vlans at layer 2 | 13:36 |
| ralonsoh | ygk_12345, you need to segment your L2 using VLAN | 13:37 |
| ralonsoh | and the routing is done outside OpenStack | 13:37 |
| pjakuszew | lajoskatona: thanks, I guess it's good idea to increase that value if I have lots of VPN connections? | 13:37 |
| ralonsoh | that's the whole point of this feature | 13:37 |
| lajoskatona | pjakuszew: for that I don't have answer, I never used vpnaas... | 13:42 |
| pjakuszew | ah ok | 13:42 |
| pjakuszew | I'm trying to debug this issue where I keep seeing following message in l3-agent log | 13:42 |
| pjakuszew | Function 'neutron_vpnaas.services.vpn.device_drivers.strongswan_ipsec.StrongSwanDriver.report_status' run outlasted interval by 12.27 sec | 13:42 |
| pjakuszew | I think that loopingcall.FixedIntervalLoopingCall is glitched somehow | 13:43 |
| pjakuszew | to be more specific, this looping call is in self.process_status_cache_check at IPsecDriver class | 13:44 |
| pjakuszew | I could just restart l3-agent, but I really want to avoid that because there are lots of routers on them and migrating them to another l3 nodes is something I always consider to be risky | 13:45 |
| pjakuszew | especially in this situation where I'm not sure that L3 agent is functioning correctly at all three nodes I have | 13:45 |
| pjakuszew | I'm seeing that loopingcall message on two of them, last one does not spam it | 13:46 |
| pjakuszew | because of that stuck loopingcall, l3-agent keeps starting neutron-vpnaas-netns-wrapper (...) --cmd ipsec,status continuously | 13:47 |
| pjakuszew | another way to go would be migrating only the routers that are hosting VPN connections, but that's something I consider to be risky too | 13:48 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: Remove ``ConfDriver`` code https://review.opendev.org/c/openstack/neutron/+/790999 | 14:05 |
| *** rpittau is now known as rpittau|afk | 14:08 | |
| opendevreview | Przemyslaw Szczerbik proposed openstack/neutron master: ovs-agent: Report pkt processing info in heartbeat https://review.opendev.org/c/openstack/neutron/+/800444 | 14:08 |
| opendevreview | Przemyslaw Szczerbik proposed openstack/neutron master: Report CUSTOM_VNIC_TYPE_ traits on OVS agent RP https://review.opendev.org/c/openstack/neutron/+/800445 | 14:08 |
| opendevreview | Przemyslaw Szczerbik proposed openstack/neutron master: Report pkt processing capacity on OVS agent RP https://review.opendev.org/c/openstack/neutron/+/800446 | 14:09 |
| opendevreview | Przemyslaw Szczerbik proposed openstack/neutron master: Bump os-resource-classes lib to 1.1.0 https://review.opendev.org/c/openstack/neutron/+/801066 | 14:09 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: Import ABC classes from collection.abc https://review.opendev.org/c/openstack/neutron/+/801068 | 14:28 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: Use explicit select condition in SQL query in "_port_filter_hook" https://review.opendev.org/c/openstack/neutron/+/801076 | 14:57 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: Add a privsep context only for link commands https://review.opendev.org/c/openstack/neutron/+/800686 | 15:10 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: [WIP] Remove ``ConfDriver`` code https://review.opendev.org/c/openstack/neutron/+/790999 | 15:22 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: Import ABC classes from collection.abc https://review.opendev.org/c/openstack/neutron/+/801068 | 15:43 |
| opendevreview | Takashi Kajinami proposed openstack/neutron-lib master: Replace deprecated import of ABCs from collections https://review.opendev.org/c/openstack/neutron-lib/+/801084 | 15:47 |
| opendevreview | Takashi Kajinami proposed openstack/neutron-lib master: Replace deprecated import of ABCs from collections https://review.opendev.org/c/openstack/neutron-lib/+/801084 | 16:22 |
| ygk_12345 | IDENTIFY | 16:31 |
| opendevreview | Merged openstack/neutron stable/ussuri: Provide integer number to arping "-w" parameter https://review.opendev.org/c/openstack/neutron/+/800868 | 17:14 |
| *** tbachman is now known as Guest1083 | 18:59 | |
| opendevreview | Merged openstack/neutron-tempest-plugin master: Enable tls-proxy support for test jobs https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/796778 | 19:05 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!