[00:43:25] *** Joins: tomzawadzki (uid327004@gateway/web/irccloud.com/x-wlkhjlfwksppwkjc) [01:02:37] *** Joins: mszwed (~mszwed@134.134.139.72) [02:16:51] *** Quits: guerby (~guerby@april/board/guerby) (Remote host closed the connection) [02:17:48] *** Joins: guerby (~guerby@april/board/guerby) [03:23:42] *** Joins: gila (~gila@5ED74129.cm-7-8b.dynamic.ziggo.nl) [08:11:47] *** Joins: vmysak (~vmysak@192.55.54.38) [08:17:41] jimharris, bwalker: https://review.gerrithub.io/c/spdk/spdk/+/432576 Revert "env: remove default pre-reserved memory size" [08:18:13] nvmf tests are failing terribly right now [08:32:00] i see fedora-03 in the chandler pool still failed with this patch - but it's the json_config.sh failure that pniedzwx alluded to earlier today [08:38:24] the patch with nvmf hooks introduced another regression [08:39:08] nvme_rdma_mr_map_notify is now associated with a single nvme_rdma_ctrlr, [08:39:50] but the entire mem_map is shared between multiple controllers [08:41:19] so nvme_rdma_mr_map_notify may try to dereference an nvme_rdma_ctrlr that was already detached and freed [08:41:35] is that the issue with json_config.sh though? [08:41:51] i don't think so [08:53:52] bwalker: jimharris: Following pull's of my repo from gerrithub, I'm still now seeing a v18.10 tag. I do see that tag, however, after I cloned the spdk from github. [08:54:03] s/now/NOT/ [09:20:13] git fetch --tags? [09:20:17] darsto: here's a commit you can try your new +2 on :) https://review.gerrithub.io/#/c/spdk/spdk.github.io/+/432499/ [09:20:37] no - lhodev is right, the gerrithub repository doesn't have the tag [09:24:48] oh [09:24:54] I guess they don't automatically sync [09:25:05] the tags at least [09:26:04] the sync is only gerrithub => github I thought [09:31:50] it looks like you are right - I just pushed the tag to gerrithub [09:36:37] *** Quits: gila (~gila@5ED74129.cm-7-8b.dynamic.ziggo.nl) (Quit: My Mac Pro has gone to sleep. ZZZzzz…) [09:39:08] darsto: taking a look at the memory region stuff now [09:52:18] bwalker: Can you help me understand precisely the sequence of events that took place with respect to the tagging (on both/either/or github and gerrithub)? And, how the sync'ing works? [09:52:40] so anything we do on GerritHub automatically syncs to GitHub [09:52:46] and we do almost everything on GerritHub [09:53:08] for a release, we do a commit that changes include/spdk/version.h to the correct version and commit it [09:53:12] which syncs to GitHub [09:53:26] then on GitHub we do a release on the GitHub webpage [09:53:38] which automatically generates a tag on whatever the latest commit on master is [09:53:39] If that's the case, then I might have expected to see the tag on GerritHub, and then later on GitHub after a sync. However, I saw the tag *first* on GitHub -- not at all on GerritHub until just now. [09:54:07] then we do another commit on GerritHub that moves the version to the next one with the "-pre" tag [09:54:24] *** Joins: gila (~gila@5ED74129.cm-7-8b.dynamic.ziggo.nl) [09:54:24] so the tag is created as part of clicking buttons on GitHub, so it stays on GitHub [09:56:04] darsto: I see bug in the hooks patch - will push fix shortly [09:58:35] all of the jenkins tests are failing on ioat_autotest [10:03:04] *** Quits: gila (~gila@5ED74129.cm-7-8b.dynamic.ziggo.nl) (Quit: My Mac Pro has gone to sleep. ZZZzzz…) [10:12:33] *** Quits: tomzawadzki (uid327004@gateway/web/irccloud.com/x-wlkhjlfwksppwkjc) (Quit: Connection closed for inactivity) [10:16:46] the ioat issue popped up just now [10:17:07] just after klateck went home :) [10:20:08] that's when the bugs always show up [10:22:48] *** Quits: vmysak (~vmysak@192.55.54.38) (Ping timeout: 268 seconds) [10:32:57] darsto do you want to hit merge on https://review.gerrithub.io/#/c/spdk/spdk.github.io/+/432499/ just to try it [10:33:47] also I think this will be the fix for the hooks problem: https://review.gerrithub.io/#/c/432590/ [10:33:48] just did it, thanks! [10:55:39] *** Joins: travis-ci (~travis-ci@ec2-54-145-14-105.compute-1.amazonaws.com) [10:55:40] (spdk/master) reduce: close backing dev during unload (Jim Harris) [10:55:40] Diff URL: https://github.com/spdk/spdk/compare/5aace13984b8...d912dba2cbab [10:55:40] *** Parts: travis-ci (~travis-ci@ec2-54-145-14-105.compute-1.amazonaws.com) () [11:28:04] jimharris: what do you want to do about this one? https://review.gerrithub.io/#/c/spdk/spdk/+/432576/ [11:28:20] it does indeed fix one of the issues that's failing every test at the moment [11:28:54] it didn't pass all tests itself for two reasons. On the chanderl pool it hit an intermittent failure in the json_config test where somehow we're getting an examine callback for the same bdev twice [11:29:01] we should merge it - i just retriggered it [11:29:15] and on the jenkins pool it still is failing ioat things [11:29:32] (you removed the verified -1 flag but that only works on the chandler pool) [11:29:41] yeah I only did that because I was just going ot hit submit [11:30:59] i'm ok merging it - peluse, sethhowe - is anyone here in chandler able to look at the ioat_autotest failure? [11:35:41] *** Joins: travis-ci (~travis-ci@ec2-54-221-104-200.compute-1.amazonaws.com) [11:35:42] (spdk/master) Revert "env: remove default pre-reserved memory size" (Darek Stojaczyk) [11:35:43] Diff URL: https://github.com/spdk/spdk/compare/d912dba2cbab...a817ccf571f1 [11:35:43] *** Parts: travis-ci (~travis-ci@ec2-54-221-104-200.compute-1.amazonaws.com) () [12:28:27] the passthru "double examine" is really two bugs [12:28:44] 1) the json clear_config.py script doesn't account for passthru [12:29:07] 2) the PT module doesn't check for duplicates [12:29:54] maybe its three bugs [12:30:13] 3) RPC completes successfully if the passthru bdev can't be created [12:39:41] how is that all hitting just now? [12:39:54] timing changes due to dynamic memory allocation? [12:40:08] the json_config failure I saw had nothing to do with the passthru issue [12:44:33] are you working on fixes or do you want me to take one? [12:44:51] pushing them now [12:45:59] only this first one is really needed - https://review.gerrithub.io/#/c/spdk/spdk/+/432600/ [12:50:36] saw your comment and repushed https://review.gerrithub.io/#/c/spdk/spdk/+/432601/ [12:50:41] bwalker ^^ [12:52:08] re-reviewed. Moved to front of queue [14:06:27] jimharris, I can (ioat_autotest failure), still need someone to look at it? [14:44:04] *** Joins: LiuXiaodong (c0373627@gateway/web/freenode/ip.192.55.54.39) [14:55:31] peluse: sethhowe found the system that was causing the ioat_autotest failures and took it offline - that test will now run on other systems [14:55:59] sethhowe: can you make sure you ping klateck about it so he looks at that system on Monday? [14:59:26] sweet [14:59:54] jimharris: will do. [15:52:20] jimharris: looks like GP3 ran a few times and then started failing as well. I have root caused it though. there is an NFS that the machines are failing to mount. I'm trying to re-enable it now. [16:14:32] cool - just saw the GP3 failure now [18:52:05] *** Quits: LiuXiaodong (c0373627@gateway/web/freenode/ip.192.55.54.39) (Ping timeout: 256 seconds)