*** dharinic has quit IRC | 00:04 | |
*** hongbin has quit IRC | 00:15 | |
*** ihrachys has joined #openstack-oslo | 00:37 | |
*** ihrachys has quit IRC | 00:37 | |
*** ansmith has quit IRC | 00:46 | |
*** tovin07_ has joined #openstack-oslo | 01:02 | |
*** ihrachys has joined #openstack-oslo | 01:04 | |
*** ihrachys has quit IRC | 01:17 | |
*** ihrachys has joined #openstack-oslo | 01:17 | |
*** zhangguoqing has joined #openstack-oslo | 01:19 | |
openstackgerrit | Merged openstack/tooz master: etcd3: add etcd3 coordination driver https://review.openstack.org/447223 | 01:20 |
---|---|---|
*** zhangguoqing has quit IRC | 01:33 | |
*** zhangguoqing has joined #openstack-oslo | 01:34 | |
*** gcb has joined #openstack-oslo | 01:39 | |
*** yamahata has quit IRC | 02:14 | |
*** eliqiao has joined #openstack-oslo | 02:48 | |
*** zhangguoqing has quit IRC | 03:01 | |
*** tonyb has quit IRC | 03:04 | |
*** dave-mccowan has quit IRC | 03:05 | |
*** nicolasbock has quit IRC | 03:05 | |
*** tonyb has joined #openstack-oslo | 03:09 | |
*** ihrachys has quit IRC | 03:09 | |
*** tonyb has quit IRC | 03:12 | |
*** tonyb has joined #openstack-oslo | 03:13 | |
*** zhangguoqing has joined #openstack-oslo | 03:13 | |
*** amotoki has joined #openstack-oslo | 03:22 | |
*** tonyb has quit IRC | 03:28 | |
*** tonyb has joined #openstack-oslo | 03:28 | |
*** links has joined #openstack-oslo | 03:40 | |
*** Dinesh_Bhor has joined #openstack-oslo | 03:44 | |
*** zhangguoqing has quit IRC | 03:45 | |
*** eliqiao has quit IRC | 03:50 | |
*** zhangguoqing has joined #openstack-oslo | 04:39 | |
*** zhangguoqing has quit IRC | 04:41 | |
*** zhangguoqing has joined #openstack-oslo | 04:57 | |
*** jamielennox is now known as jamielennox|away | 05:12 | |
*** jamielennox|away is now known as jamielennox | 05:17 | |
*** mhickey has joined #openstack-oslo | 05:26 | |
*** gcb has quit IRC | 05:33 | |
*** gcb has joined #openstack-oslo | 05:46 | |
*** rcernin has joined #openstack-oslo | 06:06 | |
*** hjensas has quit IRC | 06:18 | |
*** tesseract has joined #openstack-oslo | 06:40 | |
*** pooja_jadhav has joined #openstack-oslo | 06:46 | |
pooja_jadhav | gcb:Hi | 06:59 |
gcb | pooja_jadhav, hi | 06:59 |
*** pcaruana has joined #openstack-oslo | 06:59 | |
pooja_jadhav | gcb: I want to discuss about this bug https://bugs.launchpad.net/nova/+bug/1680130 | 06:59 |
openstack | Launchpad bug 1680130 in OpenStack Compute (nova) "Attaching a volume w/ an invalid (too long) UUID results in HTTP 500." [Undecided,In progress] - Assigned to Pooja Jadhav (poojajadhav) | 06:59 |
pooja_jadhav | gcb: If we passed invalid long volume uuid (includes hyphen) to "attach-volume" API then it returns 500 because it is trying to insert that volume uuid in to the table and volume uuid column having size 36. If more than 36 characters tries to add for volume uuid then it returns DBDataError. | 07:04 |
pooja_jadhav | gcb: so my concern we should fix this issue in the central place that is "is_uuid_like" method that removes hyphen and return true or false respectively. | 07:06 |
pooja_jadhav | we should check uuid size, if its more than 36 we should return false. | 07:06 |
*** shardy has joined #openstack-oslo | 07:09 | |
gcb | pooja_jadhav, i'm in a meeting, will check this later | 07:09 |
pooja_jadhav | gcb: Fine thanks | 07:18 |
gcb | I'm back, checking the bug ... | 07:18 |
pooja_jadhav | gcb: ok | 07:19 |
gcb | pooja_jadhav, do you mean https://github.com/openstack/oslo.utils/blob/master/oslo_utils/uuidutils.py#L45 ? | 07:23 |
pooja_jadhav | gcb: yes | 07:24 |
gcb | pooja_jadhav, valid uuid with hyphen has 36 characters, and '11111111-2222-4f9d-5555--666666666666' pass the check now ,right ? | 07:27 |
gcb | yes, we need check string's size which equals 36 | 07:28 |
pooja_jadhav | gcb: yes | 07:29 |
gcb | that makes sense , uuidutils.is_uuid_like('11111111-2222-4f9d-5555--666666666666') should return False | 07:30 |
pooja_jadhav | gcb: most of the projects uses this method to check uuid string format for validation so can we make validation of uuid in "is_uuid_like" | 07:31 |
gcb | yes, please add the length check, and I will help review | 07:33 |
gcb | or I can post the patch | 07:34 |
*** jamielennox is now known as jamielennox|away | 07:34 | |
pooja_jadhav | gcb: yes, I will fix this as early as possible.. thanks for your opinion and time :) | 07:34 |
gcb | pooja_jadhav, thanks for raising this, please add me as reviewer :-) | 07:36 |
pooja_jadhav | gcb: sure, I will add:-) | 07:38 |
*** lucas-afk is now known as lucasagomes | 07:59 | |
*** zzzeek has quit IRC | 08:00 | |
*** zzzeek has joined #openstack-oslo | 08:00 | |
*** jaosorior has joined #openstack-oslo | 08:03 | |
*** hjensas has joined #openstack-oslo | 08:22 | |
*** jamielennox|away is now known as jamielennox | 08:28 | |
*** openstackgerrit has quit IRC | 08:33 | |
*** mhickey has quit IRC | 08:34 | |
*** pblaho has joined #openstack-oslo | 08:55 | |
*** cdent has joined #openstack-oslo | 09:09 | |
*** sambetts|afk is now known as sambetts | 09:19 | |
*** e0ne has joined #openstack-oslo | 09:24 | |
*** openstackgerrit has joined #openstack-oslo | 09:37 | |
openstackgerrit | Stephen Finucane proposed openstack-dev/pbr master: trivial: Add note about multiple builders support https://review.openstack.org/439545 | 09:37 |
*** zhangguoqing has quit IRC | 09:40 | |
*** mhickey has joined #openstack-oslo | 09:49 | |
*** boden has joined #openstack-oslo | 09:58 | |
*** nicolasbock has joined #openstack-oslo | 10:04 | |
*** shardy has quit IRC | 10:06 | |
sileht | 1 | 10:08 |
*** tovin07_ has quit IRC | 10:13 | |
*** zhangguoqing has joined #openstack-oslo | 10:19 | |
*** zhangguoqing has quit IRC | 10:26 | |
*** sdague has joined #openstack-oslo | 10:26 | |
*** shardy has joined #openstack-oslo | 10:29 | |
*** zhangguoqing has joined #openstack-oslo | 10:33 | |
*** zhangguoqing has quit IRC | 10:42 | |
*** jaosorior has quit IRC | 10:55 | |
*** zhangguoqing has joined #openstack-oslo | 10:57 | |
*** geekinutah has quit IRC | 10:58 | |
*** geekinutah has joined #openstack-oslo | 10:58 | |
*** jaosorior has joined #openstack-oslo | 11:01 | |
*** shardy is now known as shardy_lunch | 11:06 | |
*** lucasagomes is now known as lucas-hungry | 11:14 | |
*** geguileo has joined #openstack-oslo | 11:18 | |
geguileo | gus: ping - oslo.privsep question | 11:24 |
geguileo | Anybody around willing to answer a privsep question? | 11:37 |
geguileo | re-reading that it sounded bad, by willing I meant that had the knowledge and the time. lol | 11:40 |
*** eck`gone is now known as eck` | 11:40 | |
*** jamielennox has left #openstack-oslo | 11:49 | |
*** jamielennox has joined #openstack-oslo | 11:49 | |
*** yassine has joined #openstack-oslo | 11:53 | |
*** yassine is now known as Guest86170 | 11:53 | |
*** Guest15696 has quit IRC | 11:54 | |
*** dave-mccowan has joined #openstack-oslo | 12:09 | |
*** shardy_lunch is now known as shardy | 12:16 | |
*** salv-orlando has joined #openstack-oslo | 12:19 | |
*** ansmith has joined #openstack-oslo | 12:22 | |
*** lucas-hungry is now known as lucasagomes | 12:24 | |
*** gordc has joined #openstack-oslo | 12:28 | |
gus | geguileo: ? | 12:32 |
geguileo | gus: Hi | 12:33 |
geguileo | gus: I would like to check, is an privsep entrypoint only able to have 1 simultatenous call? | 12:33 |
gus | "don't ask to ask, just ask" | 12:33 |
geguileo | :-) | 12:34 |
gus | (trying to remember the code ;) | 12:34 |
geguileo | thanks :-) | 12:34 |
gus | Yes, only one simultaneous call iirc. | 12:34 |
gus | multiple threads can make calls, but they get serialised by some locks. | 12:35 |
geguileo | ooooooh, that's bad :-( | 12:35 |
geguileo | Any way to work around that? | 12:35 |
* geguileo was actually hoping he had read the code wrong | 12:37 | |
gus | Make Daemon.loop fancier: https://github.com/openstack/oslo.privsep/blob/master/oslo_privsep/daemon.py#L433 | 12:38 |
gus | Currently it reads messages from the communication channel and calls functions (see _process_cmd) | 12:39 |
geguileo | I read the code, I was just hoping that I missed something or there was some kind of workaround for it | 12:39 |
geguileo | :'-( | 12:39 |
geguileo | gus thanks for the confirmation | 12:40 |
gus | You'd need to change it to use a worker thread pool or something instead of just a naive function call. The downside of course is that things that *are* just short function calls would now be slower. | 12:40 |
gus | The communication protocol has everything tagged by message id, so you can return messages out of order just fine. | 12:41 |
gus | geguileo: what long-lived thing do you need to do? | 12:42 |
geguileo | gus: well, in os-brick we do calls to iscsiadm and multipath | 12:43 |
geguileo | And if there are network issues those can take a very long time, for example a login to iSCSI when there are errors can take up to 2 minutes | 12:43 |
geguileo | to fail | 12:43 |
gus | ah :( | 12:43 |
geguileo | So that means that there can be no other call at the same time | 12:43 |
geguileo | And I had just created a patch that refactored the connect so it was done in parallel | 12:44 |
gus | yeah sorry :( | 12:44 |
geguileo | So if we wanted to create 4 sessions and 3 were down it would only take 2 minutes and not 3*2 | 12:44 |
gus | most of the hard work is done - the communication protocol supports out of order messages, and multiplexing onto the channel is protected with a lock. | 12:45 |
gus | You just need to actually invoke the privileged functions asynchronously on the server side. | 12:45 |
geguileo | I'll try to finish with the work I'm doing and see if I can poke at it | 12:46 |
geguileo | thank you for your time | 12:46 |
*** kgiusti has joined #openstack-oslo | 12:46 | |
gus | iirc I had plans to support slow generator functions too, that would be tagged with a different message type (not Message.CALL) and could dribble results back intermixed with responses from other functions. | 12:47 |
gus | that's a bit more work though (not that much). | 12:47 |
*** dimtruck is now known as zz_dimtruck | 12:49 | |
*** gcb has quit IRC | 13:00 | |
geguileo | the think is that these call can take 1 second or take 2 minutes, we don't know beforehand :-( | 13:03 |
*** dougshelley66 has left #openstack-oslo | 13:06 | |
*** jaosorior has quit IRC | 13:06 | |
*** jaosorior has joined #openstack-oslo | 13:14 | |
*** pcaruana has quit IRC | 13:27 | |
*** hjensas has quit IRC | 13:35 | |
*** links has quit IRC | 13:35 | |
*** hongbin has joined #openstack-oslo | 13:50 | |
*** pcaruana has joined #openstack-oslo | 14:10 | |
*** zz_dimtruck is now known as dimtruck | 14:11 | |
*** dave-mccowan has quit IRC | 14:24 | |
*** zhangguoqing has quit IRC | 14:36 | |
*** dave-mccowan has joined #openstack-oslo | 14:44 | |
*** efried has joined #openstack-oslo | 14:46 | |
efried | harlowja yt? | 14:47 |
*** shyama has joined #openstack-oslo | 14:52 | |
efried | harlowja - shyama discovered a TaskFlow behavior we'd like to ask you about. It reduces to something like this: https://pastebin.com/3veV7w5E | 14:53 |
efried | harlowja The behavior is that we get the exception from the revert, but the exception from the execute is lost. | 14:54 |
efried | Is this something we have to code around by explicitly swallowing exceptions in revert, and/or explicitly printing stack traces in execute? | 14:54 |
efried | Or is there some other option? | 14:54 |
*** dharinic has joined #openstack-oslo | 14:59 | |
*** ianychoi has quit IRC | 15:00 | |
*** dharinic has quit IRC | 15:04 | |
*** rcernin has quit IRC | 15:09 | |
*** Guest86170 has quit IRC | 15:21 | |
*** Guest86170 has joined #openstack-oslo | 15:25 | |
openstackgerrit | Gage Hugo proposed openstack-dev/pbr master: Fix missing comment from previous change https://review.openstack.org/457692 | 15:26 |
*** pcaruana has quit IRC | 15:41 | |
*** e0ne has quit IRC | 15:42 | |
*** salv-orl_ has joined #openstack-oslo | 15:49 | |
*** ihrachys has joined #openstack-oslo | 15:51 | |
*** salv-orlando has quit IRC | 15:51 | |
*** salv-orl_ has quit IRC | 16:14 | |
*** lucasagomes is now known as lucas-afk | 16:15 | |
*** mhickey has quit IRC | 16:30 | |
*** dave-mccowan has quit IRC | 16:41 | |
*** jaosorior has quit IRC | 16:43 | |
*** dave-mccowan has joined #openstack-oslo | 16:45 | |
*** harlowja_ has joined #openstack-oslo | 16:50 | |
*** harlowja has quit IRC | 16:52 | |
*** amotoki has quit IRC | 16:57 | |
efried | harlowja_ Can you see backscroll from ~2h15m ago? | 17:03 |
harlowja_ | ya, will get around to that in a few :) | 17:03 |
efried | harlowja_ Thanks! | 17:04 |
*** cdent has quit IRC | 17:07 | |
harlowja_ | np | 17:09 |
*** yamahata has joined #openstack-oslo | 17:09 | |
*** gnarld_ is now known as cFouts | 17:17 | |
*** shyama_ has joined #openstack-oslo | 17:18 | |
*** shyama has quit IRC | 17:20 | |
*** shyama_ is now known as shyama | 17:20 | |
*** ihrachys_ has joined #openstack-oslo | 17:23 | |
*** ihrachys has quit IRC | 17:25 | |
*** shardy has quit IRC | 17:30 | |
*** dougshelley66 has joined #openstack-oslo | 17:41 | |
openstackgerrit | Merged openstack-dev/pbr master: Fix missing comment from previous change https://review.openstack.org/457692 | 17:41 |
*** shyama has quit IRC | 17:41 | |
*** shyama has joined #openstack-oslo | 17:42 | |
openstackgerrit | Andy Smith proposed openstack/oslo.messaging master: Add get_rpc_transport call https://review.openstack.org/454194 | 17:51 |
*** nicolasbock has quit IRC | 18:01 | |
harlowja_ | efried hmmmm | 18:08 |
harlowja_ | i thought it retained both | 18:08 |
*** sambetts is now known as sambetts|afk | 18:08 | |
harlowja_ | but u might be right | 18:08 |
efried | harlowja_ It might "retain" - but doesn't print AFAICT | 18:08 |
harlowja_ | efried something to see if is logged | 18:11 |
harlowja_ | https://github.com/openstack/taskflow/blob/master/taskflow/engines/action_engine/builder.py#L244 | 18:11 |
efried | harlowja_ Would require taskflow logging at debug level, nah? | 18:12 |
harlowja_ | ya | 18:12 |
harlowja_ | also efried https://github.com/openstack/taskflow/blob/master/taskflow/engines/action_engine/engine.py#L334-L340 looks like it should capture both | 18:16 |
harlowja_ | but let's see | 18:17 |
efried | harlowja_ I tried the script in that paste with debug logging enabled (I think) and just got the one error. | 18:17 |
harlowja_ | kk | 18:17 |
efried | wanna see it? | 18:17 |
harlowja_ | nah, i got it :-P | 18:17 |
efried | k | 18:17 |
harlowja_ | ya, hmmm, ok, am thinking something not capturing an exception somewhere | 18:18 |
efried | shyama ^^ FYI | 18:19 |
*** tesseract has quit IRC | 18:22 | |
*** nicolasbock has joined #openstack-oslo | 18:27 | |
harlowja_ | efried so i think i know what to change | 18:38 |
harlowja_ | the question is what do we want to happen :) | 18:38 |
efried | Well. | 18:38 |
efried | shyama Feel free to chime in here. | 18:39 |
efried | harlowja_ I think we should log the revert exception, and raise the execute exception. | 18:39 |
efried | Question: if one revert fails, do we continue with other reverts? | 18:39 |
harlowja_ | ya, that's ^ the bigger question :-P | 18:39 |
openstackgerrit | Joshua Harlow proposed openstack/taskflow master: Allow gathering + continutation of successive REVERT_FAILURES https://review.openstack.org/457764 | 18:40 |
harlowja_ | efried ^ does continue with other reverts | 18:40 |
efried | I'd say yes. It's pretty probable they'll fail, but they should at least be given a shot. | 18:40 |
efried | Can reverts 'provide' things for other reverts? | 18:40 |
harlowja_ | nope | 18:41 |
efried | Then definitely yes, keep trying 'em. | 18:41 |
efried | harlowja_ Want me to patch that up and try it? | 18:42 |
efried | or did you already do that? | 18:43 |
harlowja_ | ya, i'm trying with little sample | 18:44 |
harlowja_ | import logging | 18:45 |
harlowja_ | logging.basicConfig(level=5) | 18:46 |
harlowja_ | btw that will show u more of the decisions being made ;) | 18:46 |
efried | harlowja_ Is LOG.isEnabledFor(DEBUG) really necessary? That logic exists in LOG.debug... oh, it's so you don't do the get_atom_intention processing huh? | 18:47 |
harlowja_ | ya | 18:47 |
efried | dig | 18:47 |
harlowja_ | which may be somewhat expensive | 18:47 |
harlowja_ | and may happen to often if left just to happen | 18:47 |
harlowja_ | and therefore suck | 18:48 |
harlowja_ | lol | 18:48 |
efried | Yuh, gotcha. | 18:48 |
efried | Oh, look, that's what your comment says. | 18:48 |
efried | Wasn't immediately obvious to me that "making any intention request to storage" == get_atom_intention() | 18:49 |
harlowja_ | ya, fair, gotta get the data from somewhere :-p | 18:49 |
harlowja_ | and if that goes out to zookeeper or something, that'd suck | 18:49 |
openstackgerrit | Joshua Harlow proposed openstack/taskflow master: Allow gathering + continutation of successive REVERT_FAILURES https://review.openstack.org/457764 | 18:51 |
harlowja_ | ok, made one small other change | 18:51 |
harlowja_ | we may want to make this an engine option | 18:51 |
harlowja_ | because its slightly hard to tell if people are depending on what exists | 18:51 |
harlowja_ | example i am running https://gist.github.com/harlowja/1151348e065c3156d62d4870cee3471e | 18:53 |
openstackgerrit | Elancheran S proposed openstack/oslo.messaging master: Retry support for oslo_messaging_notifications driver https://review.openstack.org/437673 | 18:53 |
harlowja_ | code @ https://gist.github.com/harlowja/aeec4ca758f69e9c27c821d71b00a860 | 18:53 |
harlowja_ | efried shyama ^ | 18:53 |
harlowja_ | the logging is pretty useful when needed :-p | 18:54 |
harlowja_ | but can be freaking noisy, lol | 18:54 |
harlowja_ | u now get a `taskflow.exceptions.WrappedFailure: WrappedFailure: [Failure: ValueError: foo, Failure: ValueError: bar]` | 18:55 |
harlowja_ | which is both exceptions | 18:55 |
harlowja_ | if other reverts also bork, then they will all get wrapped up | 18:56 |
*** cdent has joined #openstack-oslo | 18:58 | |
efried | harlowja_ Hm, okay. In my albeit limited experience, WrappedFailure has to be expected and dissected very carefully if I want to log the independent stack traces. | 18:58 |
harlowja_ | yup | 18:58 |
harlowja_ | eventually all that will move to https://github.com/harlowja/failure | 18:59 |
harlowja_ | along with oslo.messaging (which has similar wrapping code) | 18:59 |
harlowja_ | ^ for remote exceptions (from some RPC channel) | 18:59 |
harlowja_ | but yes, u can access the indiviudal failure from engine.storage.get_execute_failures or engine.storage.get_revert_failures | 19:00 |
harlowja_ | if u so desire | 19:00 |
efried | harlowja_ You mean with the already-existing code? | 19:00 |
harlowja_ | well revert_failures i believe are causing engine stop | 19:01 |
harlowja_ | *currently | 19:01 |
harlowja_ | so u won't get X of them | 19:02 |
harlowja_ | but u should be able to get the individual forward failure and backwards (revert) failure from those methods right now | 19:02 |
efried | harlowja_ Not having the revert failures after the first one ain't as bad as not having the execute failure. | 19:02 |
harlowja_ | ya | 19:02 |
*** JayF has left #openstack-oslo | 19:03 | |
harlowja_ | engine.storage.get_execute_failures should have the forward one right now | 19:03 |
harlowja_ | (forward == execute) | 19:03 |
efried | harlowja_ But how would I use that? | 19:04 |
efried | If I wrap my engine.run in a try/except, can I tell (without string scraping) that the exception I caught was from revert vs execute? | 19:04 |
efried | I guess I can compare them | 19:05 |
efried | try: engine.run(); except e: if e is not engine.get_execute_failures()[0]: LOG.exception(e); raise engine.get_execute_failures()[0] | 19:06 |
efried | or something. | 19:06 |
*** dave-mccowan has quit IRC | 19:06 | |
efried | Lose the original context that way, though. | 19:07 |
efried | I suppose we should really use a philosophy of "reverts shall not fail", though. | 19:09 |
efried | Reverts, if defined, should be robust enough to go through no matter what. | 19:09 |
efried | Otherwise we're begging to have incomplete cleanup. | 19:09 |
efried | shyama I think that's our answer, really. Your revert should be coded so it doesn't fail. | 19:10 |
efried | (I guess it's _my_ revert in this case) | 19:10 |
efried | harlowja_ Presumably your new proposal only creates WrappedFailure if there's more than one exception in play? | 19:16 |
efried | So if a) using linear, and b) my reverts never fail, it'll still act just like the old way. | 19:17 |
efried | Cause looking at our code, most of our reverts do indeed try/except and log (but don't raise) on the except path. | 19:18 |
efried | This one was missed. | 19:18 |
harlowja_ | efried yup | 19:21 |
harlowja_ | in general if u can not have reverts fail, that'd be super | 19:22 |
harlowja_ | cause it starts to get hairy with what do u want to do with such failures | 19:23 |
harlowja_ | same kind of question usually happens with except blocks that also raise | 19:23 |
harlowja_ | do u drop the new one, drop the old one, log both, do somethign else :-P | 19:24 |
harlowja_ | i've seen a mix | 19:24 |
harlowja_ | python 3.x has chained exception, so u'll get something akin to wrapped failures in it | 19:24 |
harlowja_ | *chained exceptions | 19:24 |
harlowja_ | * https://www.python.org/dev/peps/pep-3134/ | 19:24 |
harlowja_ | its just still hard to decide what to do with those imho :- | 19:25 |
harlowja_ | :-P | 19:25 |
harlowja_ | efried ^ | 19:25 |
harlowja_ | wrapped failures (and failure objects in general) try to approximate some of pep-3134 | 19:25 |
efried | harlowja_ Cool man. Thanks a bunch for the talk here. I think we're going to move forward with "reverts shall not fail". But please keep me in the loop about these changes. | 19:26 |
harlowja_ | sure | 19:26 |
efried | harlowja_ While we're on the subject of WrappedFailure, though... | 19:26 |
efried | It would be neat if WrappedFailure had a one-stop method to produce some kind of comprehensive message containing the stack traces for its constituent exceptions. | 19:27 |
efried | harlowja_ It would do something like this: https://github.com/powervm/pypowervm/blob/develop/pypowervm/utils/transaction.py#L788-L793 | 19:27 |
harlowja_ | ya, closest i got is looping over them and doing https://github.com/openstack/taskflow/blob/master/taskflow/types/failure.py#L427 | 19:27 |
efried | Though better, hopefully, cause we've had trouble with that. | 19:27 |
harlowja_ | ah, ya, u found that | 19:28 |
harlowja_ | kk | 19:28 |
harlowja_ | u should be able to get the tracebacks btw | 19:28 |
harlowja_ | LOG.exception afaik takes in a exc_info tuple | 19:28 |
harlowja_ | https://github.com/openstack/taskflow/blob/master/taskflow/types/failure.py#L308 | 19:28 |
harlowja_ | for fail in wfail: | 19:29 |
harlowja_ | LOG.error("Something broke", exc_info=fail.exc_info) | 19:29 |
harlowja_ | but sure i could see a method on wrapped failure being helpful there | 19:30 |
efried | fo sho | 19:30 |
*** dave-mccowan has joined #openstack-oslo | 19:32 | |
*** e0ne has joined #openstack-oslo | 19:36 | |
*** e0ne has quit IRC | 19:42 | |
*** e0ne has joined #openstack-oslo | 19:43 | |
*** david-lyle has joined #openstack-oslo | 19:53 | |
dkehn | oslo.db question: does anyone know how oslo.db can connect to two separate mysql instances from the same project, obviously assuming two session, config.database.connection and config.database.connect2? | 20:03 |
*** e0ne has quit IRC | 20:05 | |
dkehn | dims: ^^^ | 20:07 |
*** ildikov is now known as hypothermic_cat | 20:22 | |
*** dimtruck is now known as zz_dimtruck | 20:27 | |
*** ihrachys_ is now known as ihrachys | 20:30 | |
*** openstackgerrit has quit IRC | 20:33 | |
*** salv-orlando has joined #openstack-oslo | 20:45 | |
*** zz_dimtruck is now known as dimtruck | 20:48 | |
*** david-lyle has quit IRC | 20:49 | |
*** dkehn_ has joined #openstack-oslo | 20:50 | |
*** kgiusti has quit IRC | 20:59 | |
*** dkehn_ has quit IRC | 21:00 | |
*** dkehn_ has joined #openstack-oslo | 21:01 | |
*** sdague has quit IRC | 21:04 | |
*** cdent has quit IRC | 21:07 | |
*** david-lyle has joined #openstack-oslo | 21:10 | |
*** dkehn_ has quit IRC | 21:15 | |
*** salv-orl_ has joined #openstack-oslo | 21:48 | |
*** salv-orlando has quit IRC | 21:50 | |
*** david-lyle has quit IRC | 21:55 | |
*** boden has quit IRC | 22:00 | |
*** amotoki has joined #openstack-oslo | 22:13 | |
*** dkehn_ has joined #openstack-oslo | 22:43 | |
*** salv-orl_ has quit IRC | 22:48 | |
*** david-lyle has joined #openstack-oslo | 22:53 | |
*** david-lyle has quit IRC | 22:56 | |
*** ianychoi has joined #openstack-oslo | 23:14 | |
*** gordc has quit IRC | 23:21 | |
dims | dkehn : i don't think so, but rpodolyaka and zzzeek may know for sure | 23:29 |
*** d0ugal has quit IRC | 23:32 | |
dkehn | dims: rpodolyaka: zzzeek: thanks, I'm wondering if you implement an separate project.db.sqlalchemy.api.py you can get away with it? | 23:36 |
dkehn | for example | 23:36 |
*** amotoki has quit IRC | 23:37 | |
*** gordc has joined #openstack-oslo | 23:37 | |
*** gordc has quit IRC | 23:38 | |
*** d0ugal has joined #openstack-oslo | 23:41 | |
*** Nakato has quit IRC | 23:45 | |
*** Nakato has joined #openstack-oslo | 23:46 | |
*** amotoki has joined #openstack-oslo | 23:50 | |
*** hongbin has quit IRC | 23:55 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!