[00:03:17] Project autotest-nightly-failing build #325: STILL FAILING in 39 min. See https://dqtibwqq6s6ux.cloudfront.net for results. [00:08:58] *** Joins: travis-ci (~travis-ci@ec2-54-205-182-9.compute-1.amazonaws.com) [00:08:58] (spdk/master) scripts: Add a package which was used in ISA-L module in pkgdep.sh. (yidong0635) [00:08:58] Diff URL: https://github.com/spdk/spdk/compare/58d923e6ccc3...4626e2b5a5ab [00:08:58] *** Parts: travis-ci (~travis-ci@ec2-54-205-182-9.compute-1.amazonaws.com) () [07:06:26] *** Joins: tomzawadzki (uid327004@gateway/web/irccloud.com/x-xeobbmheitinwlof) [08:04:08] *** Joins: tkram01 (c0504633@gateway/web/freenode/ip.192.80.70.51) [08:06:13] Hello. This page "https://spdk.io/doc/blob.html" states that there is an example called Hello NVME Blob: hello_nvme_blob.c that shows how to create a blobstore with the nvme driver without using bdev but I can't find that file in git. Is it available? Thanks. [09:15:54] Hi @tkram01, actually you are right - there is no such example available. Going even further it seems that this part of docs was written quite long time ago. [09:15:54] Currently blobstore interacts with layer below only using bdev API (lib/blob/bdev/blob_bdev.c). This interface could be switched to another (like NVMe), but is not implemented at this time. [09:23:32] would it just be a matter of writing the functions defined in spdk_bs_dev for the nvme driver? [09:30:41] For the most part I think yes. Keeping in mind that mostly spdk_bs_dev callbacks relate to IO. Application using it would need to handle bringing up the NVMe and others. Similar to how users of blobstore (lvol, blobcli) handle finding the bdev or interactions with it. [11:27:49] *** Joins: travis-ci (~travis-ci@ec2-107-21-140-222.compute-1.amazonaws.com) [11:27:49] (spdk/master) test/ocf: wait for spdk app to finish at the end (Darek Stojaczyk) [11:27:49] Diff URL: https://github.com/spdk/spdk/compare/4626e2b5a5ab...d011196380f7 [11:27:49] *** Parts: travis-ci (~travis-ci@ec2-107-21-140-222.compute-1.amazonaws.com) () [11:28:19] *** Joins: travis-ci (~travis-ci@ec2-54-205-182-9.compute-1.amazonaws.com) [11:28:19] (spdk/master) test/vhost: Remove ramdisk and vhost disk from targetcli if script fails (Pawel Kaminski) [11:28:19] Diff URL: https://github.com/spdk/spdk/compare/d011196380f7...29ca9d63b80d [11:28:19] *** Parts: travis-ci (~travis-ci@ec2-54-205-182-9.compute-1.amazonaws.com) () [11:36:37] *** Joins: felipef (~felipef@cpc92310-cmbg19-2-0-cust421.5-4.cable.virginm.net) [11:38:08] *** Joins: travis-ci (~travis-ci@ec2-18-206-89-202.compute-1.amazonaws.com) [11:38:08] (spdk/master) lib/iscsi: Set destructed at the start of target node destruction (Shuhei Matsumoto) [11:38:08] Diff URL: https://github.com/spdk/spdk/compare/29ca9d63b80d...8cc7b0bc4d1a [11:38:08] *** Parts: travis-ci (~travis-ci@ec2-18-206-89-202.compute-1.amazonaws.com) () [11:41:30] *** Quits: felipef (~felipef@cpc92310-cmbg19-2-0-cust421.5-4.cable.virginm.net) (Ping timeout: 264 seconds) [13:40:11] *** Quits: tkram01 (c0504633@gateway/web/freenode/ip.192.80.70.51) (Quit: Page closed) [14:28:02] *** Joins: martinkennelly (~martinken@192.198.151.44) [14:29:24] Hey, I want to this spdk driver for cinder. I have one rdma enabled nic. I want to setup a initiator and target node. Can I use soft ROCE on one side? [14:41:15] *** Quits: martinkennelly (~martinken@192.198.151.44) (Ping timeout: 252 seconds) [14:44:19] *** Joins: Se-young (a57c9e3a@gateway/web/freenode/ip.165.124.158.58) [14:46:09] Hello all. I'd like to get some help regarding the nvmeof-tcp on spdk. Am I in the right place? [14:51:45] yep [14:53:25] @bwalker, Thank you. I've been using nvmeof over rdma for a while using spdk and want to try the tcp transport. [14:55:46] My setup involves a target which runs spdk and the initiator uses nvme-cli 1.8 on linux kernel 5.1rc4. I tried to connect to the target using the nvme-tcp kernel module but no luck. [14:56:18] what's the error message you're getting? [14:57:56] The error message is simply "Killed", then If i try the nvme connect again it freezes [14:58:18] on the initiator side, do you see anything appear in dmesg? [14:58:28] The kernel message is showing [ 166.929873] nvme nvme16: creating 5 I/O queues. [ 166.979206] BUG: unable to handle kernel NULL pointer dereference at 0000000000000190 [14:59:19] [ 166.979211] PGD 0 P4D 0 [ 166.979213] Oops: 0000 [#1] SMP PTI [ 166.979215] CPU: 1 PID: 7783 Comm: nvme Not tainted 5.1.0-050100rc4-generic #201904072131 [14:59:25] Then it continues with RIP, codes and traces [15:01:22] does the linux kernel initiator work as expected when connecting to the linux kernel target? [15:01:55] I have not tried it yet. I will try and see if it works. [15:02:35] that's probably the best place to start - switch to the kernel target or switch to the spdk initiator [15:03:24] Thank you @bwalker [15:27:51] *** Joins: felipef (~felipef@cpc92310-cmbg19-2-0-cust421.5-4.cable.virginm.net) [15:29:27] *** Quits: felipef (~felipef@cpc92310-cmbg19-2-0-cust421.5-4.cable.virginm.net) (Remote host closed the connection) [15:30:01] *** Joins: felipef (~felipef@cpc92310-cmbg19-2-0-cust421.5-4.cable.virginm.net) [15:34:21] *** Quits: felipef (~felipef@cpc92310-cmbg19-2-0-cust421.5-4.cable.virginm.net) (Ping timeout: 252 seconds) [15:34:58] I've just tried the same initiator against nvmet-tcp using the same 5.1rc4 and it seems to work OK. [15:35:25] However the spdk target does not seem to work, still. [15:40:03] If it helps, here is the spdk target config, https://pastebin.com/BmnG8SWD, and the initiator's kernel message https://pastebin.com/6GmcX5TG [16:04:55] *** Quits: Se-young (a57c9e3a@gateway/web/freenode/ip.165.124.158.58) (Ping timeout: 256 seconds) [16:40:19] *** Joins: Se-young (~se_young@dhcp-165-124-158-58.nwu.icair.org) [16:48:24] *** Quits: Se-young (~se_young@dhcp-165-124-158-58.nwu.icair.org) (Remote host closed the connection) [17:49:25] *** Joins: Leeleo3x (8a337d08@gateway/web/freenode/ip.138.51.125.8) [17:53:17] Hi all, I'm new to SPDK and have a few questions. I'm trying to use BDev on top of an NVMe device and my program may issue multiple write requests. Is the order of write requests guaranteed by SPDK? [17:56:31] Assume the function calls spdk_bdev_write_blocks multiple times in the same thread. [18:31:20] *** Joins: travis-ci (~travis-ci@ec2-54-145-210-121.compute-1.amazonaws.com) [18:31:20] (spdk/master) UT/nvmf: add nvmf reservation notification unit tests (Changpeng Liu) [18:31:20] Diff URL: https://github.com/spdk/spdk/compare/8cc7b0bc4d1a...b62dca93082b [18:31:20] *** Parts: travis-ci (~travis-ci@ec2-54-145-210-121.compute-1.amazonaws.com) () [18:36:32] *** Quits: tomzawadzki (uid327004@gateway/web/irccloud.com/x-xeobbmheitinwlof) (Quit: Connection closed for inactivity) [19:30:38] *** Joins: felipef (~felipef@cpc92310-cmbg19-2-0-cust421.5-4.cable.virginm.net) [19:34:47] *** Quits: felipef (~felipef@cpc92310-cmbg19-2-0-cust421.5-4.cable.virginm.net) (Ping timeout: 240 seconds) [23:31:18] *** Joins: felipef (~felipef@cpc92310-cmbg19-2-0-cust421.5-4.cable.virginm.net) [23:35:29] *** Quits: felipef (~felipef@cpc92310-cmbg19-2-0-cust421.5-4.cable.virginm.net) (Ping timeout: 252 seconds)