| 01:30.08 | *** join/#asterisk-dev pchero (~pchero@211.178.226.108) |
| 01:30.08 | *** join/#asterisk-dev Deeewayne (~dwayne@2605:a601:a816:5000:f7ce:535c:c78:9ec3) |
| 01:30.08 | *** join/#asterisk-dev DanFromUK (sid21651@gateway/web/irccloud.com/x-rwgvoygihokyhnsm) |
| 01:30.08 | *** mode/#asterisk-dev [+oo pchero Deeewayne] by verne.freenode.net |
| 03:32.29 | *** join/#asterisk-dev spatel (~spatel@pool-96-237-230-175.bstnma.fios.verizon.net) |
| 03:42.43 | *** join/#asterisk-dev electronic_eel_ (~quassel@213.240.182.194) |
| 07:14.33 | *** join/#asterisk-dev jkroon (~jkroon@165.16.203.60) |
| 07:14.33 | *** mode/#asterisk-dev [+o jkroon] by ChanServ |
| 08:18.24 | *** join/#asterisk-dev electronic_eel (~quassel@213.240.182.194) |
| 10:15.06 | *** join/#asterisk-dev AsteriskRoss (~AsteriskR@37.157.48.2) |
| 10:30.10 | *** join/#asterisk-dev pchero (~pchero@211.178.226.108) |
| 10:30.10 | *** mode/#asterisk-dev [+o pchero] by ChanServ |
| 12:30.22 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01:e0a:d:1111:9022:4fda:d065:8573) |
| 12:37.09 | *** join/#asterisk-dev CELYA_1 (~Thunderbi@lstlambert-657-1-90-73.w80-13.abo.wanadoo.fr) |
| 13:09.50 | *** join/#asterisk-dev spatel (~spatel@pool-96-237-230-175.bstnma.fios.verizon.net) |
| 15:02.25 | *** join/#asterisk-dev kharwell (uid358942@gateway/web/irccloud.com/x-anpfxzyliajlvxcj) |
| 15:02.25 | *** mode/#asterisk-dev [+o kharwell] by ChanServ |
| 15:12.33 | *** join/#asterisk-dev AsteriskRoss (~AsteriskR@37.157.48.2) |
| 15:21.21 | *** join/#asterisk-dev bford (uid283514@gateway/web/irccloud.com/x-sqnfevteqbrdhjvj) |
| 15:21.21 | *** mode/#asterisk-dev [+o bford] by ChanServ |
| 15:58.57 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d00dd6c39e4e3d4e964.ipv6.abo.wanadoo.fr) |
| 18:51.06 | *** join/#asterisk-dev jkroon (~jkroon@165.16.203.106) |
| 18:51.06 | *** mode/#asterisk-dev [+o jkroon] by ChanServ |
| 19:16.06 | seanbright | so for anyone looking at NULL-snapshot-adjacent crashes |
| 19:16.25 | seanbright | admitted i don't grok the stasis message bus, but |
| 19:17.15 | seanbright | it seems to me that messages are published to a queue that at some point in the future is handled and transformed |
| 19:17.44 | seanbright | for example, publishing something that contains a channel UID and then later trying to get a snapshot of that channel |
| 19:18.14 | seanbright | the channel may exist when the channel is published but not when we are generating (for example) an AMI event |
| 19:18.24 | seanbright | s/when the channel/when the message/ |
| 19:18.43 | seanbright | so a NULL snapshot is not unsurprising |
| 19:19.14 | seanbright | i assume this will happen more under load, because the time between publishing the message and handling it will increase as the task processors that are doing the pulishing fall further behind |
| 19:19.40 | kharwell | UID gets copied doesn't it into the snapshot? Same for other fields. So if the channel goes away the data is still held by the snapshot I believe |
| 19:20.03 | seanbright | there are a handful of open issues i have seen that involve ast_channel_prefix_by_snapshot(...) |
| 19:20.30 | seanbright | let me pull an example |
| 19:20.46 | seanbright | well, this is an example |
| 19:20.48 | seanbright | https://gerrit.asterisk.org/c/asterisk/+/14889 |
| 19:21.36 | file | depends on what the message contains and the implementation, many attach a snapshot at the creation of the message versus retrieving it later |
| 19:21.50 | file | also depends on when and where the message is... created |
| 19:21.57 | seanbright | actually, that is not a good example, sorry |
| 19:22.35 | seanbright | i have seen examples where we try to get the snapshot later and expect it to be non-NULL |
| 19:22.48 | seanbright | but i don't see why we would ever have that expectation |
| 19:23.11 | seanbright | maybe i am talking out of my ass |
| 19:23.16 | seanbright | let me find an example |
| 19:23.57 | kharwell | from what I can tell it's: retrieve a snapshot, attach to payload, publish to stasis, and then stasis message bus invokes callbacks with given message containing snapshot |
| 19:24.09 | kharwell | snapshot would not ever be NULL |
| 19:24.32 | kharwell | that said I am trying to find a case where the snapshot appears to be corrupted. Or previously freed |
| 19:24.40 | kharwell | so bug(s) someplace |
| 19:27.31 | seanbright | well this is embarassing |
| 19:28.15 | file | no, this is Thursday |
| 19:28.44 | seanbright | i swear i have seen cases of what i just described... maybe i am just losing my mind |
| 19:29.43 | seanbright | slowly backs out of the room... |
| 19:29.50 | seanbright | thank you for your attention |
| 19:29.53 | kharwell | some stuff changed things in 18. Maybe saw something in earlier version? |
| 19:30.04 | seanbright | perhaps |
| 19:30.14 | seanbright | that must be what it was... thanks for the life line |
| 19:30.22 | kharwell | anytime |
| 20:56.52 | *** join/#asterisk-dev AsteriskRoss (~AsteriskR@37.157.48.2) |