[00:14:42] *** Joins: klateck (klateck@nat/intel/x-lfyibqthvimhvsxd) [01:03:16] *** Joins: pwodkowx (~pwodkowx@134.134.139.72) [01:05:16] darsto: "nbd.c: 928:spdk_nbd_start: *ERROR*: ..." yup, it started showing up two or three weeks ago. [01:24:36] ohhh, looks like folks started using start_nbd_disk RPC but they forgot about "waitfornbd" function in common :D [02:16:09] *** Joins: tomzawadzki (uid327004@gateway/web/irccloud.com/x-vlwnhvrtmxwbmpfq) [02:22:24] *** Joins: pniedzwx (pniedzwx@nat/intel/x-ceumkmnbcdezxqfr) [03:43:25] *** Joins: travis-ci (~travis-ci@ec2-54-211-65-230.compute-1.amazonaws.com) [03:43:26] (spdk/master) event: Use a single event pool instead of one per socket (Ben Walker) [03:43:27] Diff URL: https://github.com/spdk/spdk/compare/e956be96eb63...e1ec5c60cae2 [03:43:27] *** Parts: travis-ci (~travis-ci@ec2-54-211-65-230.compute-1.amazonaws.com) () [05:27:50] *** Joins: felipef (~felipef@cpc92310-cmbg19-2-0-cust421.5-4.cable.virginm.net) [07:34:11] jimharris, FYI note ready yet to integrate reduce, next week.. debugging issues w/compressdev i/f and getting the mbufs setup correctly which is obviously required either way, Shouldn't take long, just don't have a ton of time this week. Will knock out some little stuff instead.... [07:58:29] jimharris, bwalker FYI compression dependency patch update https://review.gerrithub.io/c/spdk/spdk/+/429523 -- I'm outta here, have a great day! [08:01:57] *** Joins: travis-ci (~travis-ci@ec2-54-204-173-75.compute-1.amazonaws.com) [08:01:58] (spdk/master) ut_mock: rename library from spdk_mock to ut_mock (Jim Harris) [08:01:58] Diff URL: https://github.com/spdk/spdk/compare/b0281115b5a3...f425b985138d [08:01:58] *** Parts: travis-ci (~travis-ci@ec2-54-204-173-75.compute-1.amazonaws.com) () [08:05:28] *** Joins: travis-ci (~travis-ci@ec2-54-196-132-170.compute-1.amazonaws.com) [08:05:29] (spdk/master) lvol: check spdk_lvol for NULL before dereferencing it (Tomasz Zawadzki) [08:05:29] Diff URL: https://github.com/spdk/spdk/compare/f425b985138d...6bba6fe9b56d [08:05:29] *** Parts: travis-ci (~travis-ci@ec2-54-196-132-170.compute-1.amazonaws.com) () [08:06:58] *** Joins: travis-ci (~travis-ci@ec2-54-82-98-40.compute-1.amazonaws.com) [08:06:59] (spdk/master) lib/thread: check for NULL return on _get_thread() (Tomasz Zawadzki) [08:06:59] Diff URL: https://github.com/spdk/spdk/compare/6bba6fe9b56d...c49786722a15 [08:06:59] *** Parts: travis-ci (~travis-ci@ec2-54-82-98-40.compute-1.amazonaws.com) () [08:08:08] *** Joins: travis-ci (~travis-ci@ec2-54-197-136-96.compute-1.amazonaws.com) [08:08:09] (spdk/master) reactor: handle failure when getting reactor in spdk_reactor_get_tsc_stats() (Tomasz Zawadzki) [08:08:10] Diff URL: https://github.com/spdk/spdk/compare/c49786722a15...63bafd40a7ea [08:08:10] *** Parts: travis-ci (~travis-ci@ec2-54-197-136-96.compute-1.amazonaws.com) () [08:09:19] *** Joins: travis-ci (~travis-ci@ec2-54-204-169-237.compute-1.amazonaws.com) [08:09:20] (spdk/master) nvme_pcie: re-probe the pci bus on hotplug event (Darek Stojaczyk) [08:09:20] Diff URL: https://github.com/spdk/spdk/compare/63bafd40a7ea...8ee5ca50e7c1 [08:09:20] *** Parts: travis-ci (~travis-ci@ec2-54-204-169-237.compute-1.amazonaws.com) () [08:17:11] *** Joins: travis-ci (~travis-ci@ec2-54-158-193-237.compute-1.amazonaws.com) [08:17:12] (spdk/master) test/blobfs: Simplify threading in unit test (Ben Walker) [08:17:12] Diff URL: https://github.com/spdk/spdk/compare/8ee5ca50e7c1...66e620d8c56c [08:17:12] *** Parts: travis-ci (~travis-ci@ec2-54-158-193-237.compute-1.amazonaws.com) () [08:38:03] jimharris: about [08:38:31] klocwork issues [08:38:37] like https://review.gerrithub.io/#/c/spdk/spdk/+/433834/ [08:38:58] sure - whats up? [08:39:13] I think, returning nonsense is not an solution here [08:39:38] and there. I tracked some Tomek's patches and downwoted them [08:40:05] for this one we should faile RPC request or return JSON Null [08:40:18] i'm fine with that [08:43:00] the same is true for other places when klocwork find a potential bug's. Instead of masking issues we should make sure that bogus paths are impossible to take and add proper asserts there. [08:44:07] do you have links for other klocwork issues like this? [08:44:54] https://review.gerrithub.io/#/c/spdk/spdk/+/434061/ [08:45:40] this tool is usefull but also producing a lot of false positives [09:10:45] *** Joins: gila (~gila@5ED74129.cm-7-8b.dynamic.ziggo.nl) [11:09:48] *** Joins: LiuXiaodong (c037362c@gateway/web/freenode/ip.192.55.54.44) [11:13:16] pwodkowx: i agree - we need to be judicious in how we resolve these klocwork issues [11:13:39] looks like there was a merge conflict on master - i'm resolving it now [11:20:12] https://review.gerrithub.io/#/c/spdk/spdk/+/434100/ [11:20:23] bwalker: can you take a look? this fixes the build errors on master [12:15:56] *** Quits: tomzawadzki (uid327004@gateway/web/irccloud.com/x-vlwnhvrtmxwbmpfq) (Quit: Connection closed for inactivity) [12:29:14] i see a whole slew of "modification time xxx s in the future" messages on one of the ubuntu vms in jenkins: [12:29:15] https://ci.spdk.io/spdk-jenkins/results/autotest-per-patch/builds/15474/archive/ubuntu16_autotest/build.log [12:39:44] jimharris: do you want me to just merge this fix? [12:39:56] no - there's still a bug in it [12:40:00] i just pushed a new version [12:40:07] I see now that it isn't passing the tests yet [12:40:48] now that the whole-archive change is in, there's a ton of simplification we can do around our linker args (which i think would help avoid this problem in the future) [12:41:37] is the jenkins pool back up and accepting retriggers? [12:56:36] ugh - this needs some more work - i'm just going to revert my whole-archive patch for now [13:37:01] *** Joins: travis-ci (~travis-ci@ec2-54-81-15-122.compute-1.amazonaws.com) [13:37:02] (spdk/master) Revert "build: link all static libs with --whole-archive" (Jim Harris) [13:37:02] Diff URL: https://github.com/spdk/spdk/compare/66e620d8c56c...880a1426f2a3 [13:37:02] *** Parts: travis-ci (~travis-ci@ec2-54-81-15-122.compute-1.amazonaws.com) () [13:55:02] bwalker: you there? [14:10:53] yep [14:11:25] can we have this merged? https://review.gerrithub.io/c/spdk/spdk/+/434040 [14:13:02] that looks fine, but I'm not sure that's the root cause [14:14:36] I think this series addresses the root cause, which is a race between the modprobe nbd and the start_nbd_disk rpc: https://review.gerrithub.io/#/c/spdk/spdk/+/433939/ [14:15:00] that's only exposed because we start up so much faster now that dynamic memory management has been enabled [14:15:31] oh certainly, that series looks nice [14:15:54] unfortunately it seems to have hit the other main intermittent failure in nvmf a bunch of times [14:16:11] there's another patch from pwodkowx that addresses the same issue though [14:16:12] between these two intermittent failures its real hard to get a patch through [14:16:34] which one is pwodkowx patch? [14:17:26] here: https://review.gerrithub.io/c/spdk/spdk/+/413828 [14:19:45] apparently the nbd in linux is still unreliable and we need to do issue some real I/O to check if it the device is up [14:20:11] hmm [14:20:14] I'm looking this over [14:20:27] in the test pool, the intermittent failure occurs when the ioctl NBD_SET_SOCK returns EBUSY [14:20:38] I looked through the kernel code and confirmed that's the path that is problematic [14:20:46] it prints in dmesg [14:21:03] so pwodkowx's patch won't address that problem - it will still fail [14:21:35] now there may be additional problems beyond that [14:21:43] where we can't really complete that RPC until a real I/O is sent [14:22:45] I don't really understand why this patch is implemented with so many pthreads [15:19:03] *** Quits: LiuXiaodong (c037362c@gateway/web/freenode/ip.192.55.54.44) (Ping timeout: 256 seconds) [15:44:51] *** Joins: travis-ci (~travis-ci@ec2-54-81-15-122.compute-1.amazonaws.com) [15:44:52] (spdk/master) app/trace: do not print shm_name when using file_name (Tomasz Zawadzki) [15:44:53] Diff URL: https://github.com/spdk/spdk/compare/880a1426f2a3...a7e049e7226a [15:44:53] *** Parts: travis-ci (~travis-ci@ec2-54-81-15-122.compute-1.amazonaws.com) () [15:46:58] *** Joins: travis-ci (~travis-ci@ec2-54-159-180-104.compute-1.amazonaws.com) [15:46:59] (spdk/master) nvmf/host/perf.sh: refactor the code into a common function (Ziye Yang) [15:47:00] Diff URL: https://github.com/spdk/spdk/compare/a7e049e7226a...c42bee0cb6f6 [15:47:00] *** Parts: travis-ci (~travis-ci@ec2-54-159-180-104.compute-1.amazonaws.com) () [15:52:53] *** Joins: travis-ci (~travis-ci@ec2-54-204-173-75.compute-1.amazonaws.com) [15:52:54] (spdk/master) nbd: Add a function to get the nbd path (Ben Walker) [15:52:55] Diff URL: https://github.com/spdk/spdk/compare/c42bee0cb6f6...bc7b547368e1 [15:52:55] *** Parts: travis-ci (~travis-ci@ec2-54-204-173-75.compute-1.amazonaws.com) () [17:59:29] *** Joins: travis-ci (~travis-ci@ec2-54-205-30-233.compute-1.amazonaws.com) [17:59:30] (spdk/master) iscsi: check the length of initiator/target name (Changpeng Liu) [17:59:30] Diff URL: https://github.com/spdk/spdk/compare/af034b6de429...e8b65232802c [17:59:30] *** Parts: travis-ci (~travis-ci@ec2-54-205-30-233.compute-1.amazonaws.com) () [18:04:32] *** Joins: travis-ci (~travis-ci@ec2-54-82-98-40.compute-1.amazonaws.com) [18:04:33] (spdk/master) test/blobfs: Simplify threading in sync unit test (Ben Walker) [18:04:34] Diff URL: https://github.com/spdk/spdk/compare/e8b65232802c...7f6068f04d1f [18:04:34] *** Parts: travis-ci (~travis-ci@ec2-54-82-98-40.compute-1.amazonaws.com) () [18:57:49] *** Joins: travis-ci (~travis-ci@ec2-54-83-92-84.compute-1.amazonaws.com) [18:57:50] (spdk/master) vhost: Move mutex initialization after possible failure paths (Maciej Szwed) [18:57:51] Diff URL: https://github.com/spdk/spdk/compare/65724c3f882a...b86e1be001b6 [18:57:51] *** Parts: travis-ci (~travis-ci@ec2-54-83-92-84.compute-1.amazonaws.com) () [19:02:42] darsto: i noticed these first two patches aren't getting run by Jenkins: https://review.gerrithub.io/#/c/spdk/spdk/+/433863/ [19:46:20] *** Quits: felipef (~felipef@cpc92310-cmbg19-2-0-cust421.5-4.cable.virginm.net) (Read error: Connection reset by peer)