| tkajinam | hberaud, I'm attending the summit next week and I think stephenfin will be there. If he has some time then we can probably gather there to discuss that http based rpc raised in the ml | 07:53 |
|---|---|---|
| tkajinam | we probably have to see the spec doc to agree with implementation details but we can discuss the plan to add it (like test plan) | 07:54 |
| hberaud | tkajinam: I won't travel to the summit | 07:54 |
| hberaud | but yeah this is a good idea | 07:54 |
| tkajinam | hberaud, yeah I know you won't be there and I think we need some more involved later to get the spec merged. the intention of in-person discussion is more about some early discussion and feedback. | 07:55 |
| hberaud | WFM | 07:56 |
| tkajinam | I can join it alone but I want to have someone else if possible as I might miss some core items from the whole project's pov, so having at least the 2nd person would be nice. | 07:58 |
| tkajinam | If you know anyone else attending and possibly interested in this topic then we can invite him/her as well. | 07:58 |
| hberaud | good question, may Sean Mooney could be interested | 08:01 |
| tkajinam | ok | 08:02 |
| hberaud | or John Eckersberg but I don't think he will travel to the PTG too | 08:02 |
| tkajinam | yeah seems he is not in the attendee list | 08:02 |
| tkajinam | I'll wait for Stephen and check if he is available and agree with the above. We anyway need to follow up the discussion online later considering a small small amount of people attending so they can join the discussion after the summit. | 08:04 |
| hberaud | sure | 08:04 |
| hberaud | thanks for raising this topic | 08:05 |
| sahid | o/ | 08:13 |
| sahid | tkajinam: are you good with https://review.opendev.org/c/openstack/osprofiler/+/879318 ? | 08:14 |
| sahid | We have something that I need to check with you, it's related to keystonemiddleware | 08:17 |
| sahid | https://github.com/openstack/keystonemiddleware/blob/master/keystonemiddleware/auth_token/_memcache_crypt.py#L208 | 08:17 |
| sahid | expecially here ^ | 08:17 |
| sahid | in some situation we have noticed timeout in our env and the token does not get well inserted in memcached | 08:18 |
| sahid | if this happens for a service token, that means for every requests the service is doing that exception get raised and the process is terminated with error | 08:18 |
| sahid | there are no chance for the cleint to regenerate a new token until it expires | 08:19 |
| tkajinam | sahid, +2+Aed. I saw the comment from Herve so did not vote on it but it seems he already approved the change so we can move it forward | 08:20 |
| sahid | does that makes sense for you that - instead od raising an exception, we could log a warning and return like there was nothing in memcached (no token) | 08:20 |
| sahid | the client will be able to regerated a new one transparently | 08:20 |
| tkajinam | and checking that logic. give me some time. | 08:21 |
| sahid | thank you tkajinam | 08:22 |
| sahid | I will try to share some traces | 08:22 |
| tkajinam | sahid, I agree it makes sense to ignore the data if that is not decryptable but I think the current logic already ignores the error and returns None | 08:26 |
| tkajinam | https://github.com/openstack/keystonemiddleware/blob/master/keystonemiddleware/auth_token/_cache.py#L296-L297 | 08:26 |
| tkajinam | if I read the code correctly then the process should not stop immediately but the exception is logged then ignored. | 08:28 |
| sahid | https://etherpad.opendev.org/p/issue-keystone | 08:29 |
| sahid | oh... yes that is interesting | 08:29 |
| sahid | let me double check | 08:29 |
| sahid | you can see in the link the trace when we set the token | 08:29 |
| sahid | and the one we can after that | 08:29 |
| sahid | s/can/get | 08:30 |
| tkajinam | seems it is failing at a different point ? | 08:33 |
| tkajinam | https://github.com/openstack/keystonemiddleware/blob/master/keystonemiddleware/auth_token/_cache.py#L243-L244 | 08:33 |
| tkajinam | we have to ensure data is not None | 08:33 |
| tkajinam | Give me a few minutes. I can create one by myself | 08:34 |
| sahid | ah yes right :-) | 08:34 |
| sahid | cool, thanks a lot | 08:35 |
| tkajinam | sahid, do you mind sharing the versions you are using in your env ? You can share that in dm if you prefer it | 08:36 |
| sahid | tkajinam: I just shared to you in dm | 08:38 |
| stephenfin | tkajinam, hberaud: Sorry, bank holiday here yesterday. I'll be there, yes. Happy to discuss the HTTP RPC driver with the LINE folks. I wonder if we want to set up an oslo room if there are some people there? | 09:02 |
| tkajinam | stephenfin, I hope you enjoyed the long weekend. we can request adding oslo to the project lists and reserve a table (I see multiple tables are still available so we can be ad-hoc https://ptg.opendev.org/ptg.html ). | 09:37 |
| tkajinam | Probably we can create an etherpad and ask anyone interested in bringing other topics | 09:37 |
| tkajinam | I've reserved slots for heat/puppet on Wednesday and will attend these but will be free on Thursday (just one session I want to attend). I can be flexible so can get some time on Wednesday as well | 09:39 |
| sahid | tkajinam: https://review.opendev.org/c/openstack/keystonemiddleware/+/885351 | 09:51 |
| sahid | i have added you as co author, thanks a lot for your great help :-) | 09:51 |
| opendevreview | Merged openstack/osprofiler master: profiler: add python requests profile https://review.opendev.org/c/openstack/osprofiler/+/879318 | 09:59 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!