[00:09:19] Project autotest-nightly-failing build #301: STILL FAILING in 45 min. See https://ci.spdk.io/spdk-jenkins for results. [04:46:24] *** Joins: gila (~gila@134.19.189.78) [05:59:27] *** Joins: travis-ci (~travis-ci@ec2-3-89-75-137.compute-1.amazonaws.com) [05:59:28] (spdk/master) test scripts: add rpm-build package in vm_setup.sh (yidong0635) [05:59:29] Diff URL: https://github.com/spdk/spdk/compare/c2a53da73c6b...d47468a151b2 [05:59:29] *** Parts: travis-ci (~travis-ci@ec2-3-89-75-137.compute-1.amazonaws.com) () [08:06:58] *** Joins: peluse (peluse@nat/intel/x-fftlnhcliznbkppp) [08:06:58] *** ChanServ sets mode: +o peluse [09:39:14] *** Joins: travis-ci (~travis-ci@ec2-3-89-233-160.compute-1.amazonaws.com) [09:39:15] (spdk/master) test/ftl: Change fio_plugin from basic test to bdevperf (Pawel Niedzwiecki) [09:39:15] Diff URL: https://github.com/spdk/spdk/compare/d058328e12f4...ac2dd4b2c291 [09:39:15] *** Parts: travis-ci (~travis-ci@ec2-3-89-233-160.compute-1.amazonaws.com) () [10:38:46] https://review.gerrithub.io/c/spdk/spdk/+/447969/5 [10:46:16] *** Joins: travis-ci (~travis-ci@ec2-54-226-130-111.compute-1.amazonaws.com) [10:46:17] (spdk/master) bdev/reduce: fix issue with vol_load (paul luse) [10:46:17] Diff URL: https://github.com/spdk/spdk/compare/07042bdeb1d8...cf467578f89e [10:46:17] *** Parts: travis-ci (~travis-ci@ec2-54-226-130-111.compute-1.amazonaws.com) () [11:15:05] deprecating spdk_dma_malloc/free... [11:16:03] one of the problems we identified when we first looked at this, is that if an alternate environment (different than DPDK) needed to know the flags used when allocating the buffer before freeing it, how would they do that? [11:29:25] ubuntu-16 vm seems to be failing everything - fragmented huge pages [11:30:10] peluse: can you take a look? [11:30:25] or sethhowe? [12:19:27] yeah that ubuntu-16 vm has the test pool effectively down [12:35:20] jimharris: looking at it [13:44:51] *** Quits: tomzawadzki (uid327004@gateway/web/irccloud.com/x-dvccpfqulkskrcql) (Quit: Connection closed for inactivity) [14:12:50] sethhowe, thanks [14:18:30] thanks to Karol. I didn't figure it out. He logged in and worked it out. He said they were doing something with allocating hugepages for the VMs at boot and he had to change something in the VM configuration. [14:21:40] jimharris, FYI I believe I have all of your comments addressed in the base vbdev patch now. Note that the channel stuff will likely tweak on up the chain as some of it can't be tested until the tip but the tip is working right now w/the base changes. Just need to rebase the whole chain.... https://review.gerrithub.io/c/spdk/spdk/+/435747 [14:54:27] & with the top I can run bdevperf (no DPDK yet of course) for QD 128, R, W or mix all OK [14:54:35] ok - should i wait to take a look until you've rebased the whole chain? [14:58:06] its rebased now [14:58:41] the one at the very end is WIP so don't look at that one, that's DPDK integration which will take a bit mroe time. Before that is all good w/o compressdev [15:13:16] jimharris, hold off for now, there's one LOC that didn't merge right, need to manually each patch in the series.... grrr [15:44:37] *** Quits: gila (~gila@134.19.189.78) (Quit: My Mac Pro has gone to sleep. ZZZzzz…) [16:37:07] *** Joins: travis-ci (~travis-ci@ec2-54-160-226-174.compute-1.amazonaws.com) [16:37:08] (spdk/master) rdma: change structure of drained_qpair to work w/ messages. (Seth Howell) [16:37:08] Diff URL: https://github.com/spdk/spdk/compare/cf467578f89e...33668b22548b [16:37:08] *** Parts: travis-ci (~travis-ci@ec2-54-160-226-174.compute-1.amazonaws.com) () [16:43:13] *** Joins: travis-ci (~travis-ci@ec2-54-226-130-111.compute-1.amazonaws.com) [16:43:14] (spdk/master) bdev/nvme: make construct nvme device RPC can be executed asynchronously (Changpeng Liu) [16:43:15] Diff URL: https://github.com/spdk/spdk/compare/33668b22548b...e208be6f1d52 [16:43:15] *** Parts: travis-ci (~travis-ci@ec2-54-226-130-111.compute-1.amazonaws.com) () [18:19:16] jimharris, OMG the compress chain is a mess, when I rebased using the GH GUI I realized much later on that some of the in between patches were not defaulted to latest so a few of these rebases used old sets. Ugh, not sure why GH sometimes defaults to an older set, I have my screen zoomed in to much that the note indicating its not the latest is just below my scroll. Will fix tomorrow, oh what joy :) [21:06:55] *** Joins: travis-ci (~travis-ci@ec2-18-206-124-111.compute-1.amazonaws.com) [21:06:56] (spdk/master) virtio/user: fix memory hotplug with the internal rte_vhost lib (Darek Stojaczyk) [21:06:57] Diff URL: https://github.com/spdk/spdk/compare/e208be6f1d52...e6c2ff37e524 [21:06:57] *** Parts: travis-ci (~travis-ci@ec2-18-206-124-111.compute-1.amazonaws.com) () [23:38:43] *** Joins: gila (~gila@5ED4D979.cm-7-5d.dynamic.ziggo.nl)