[00:00:00] (spdk/master) doc/jsonrpc: fix up nvmf_subsystem_remove_ns (Daniel Verkamp) [00:00:00] Diff URL: https://github.com/spdk/spdk/compare/8f1c45b1b7e0...74f0dfd9b532 [00:00:00] *** Parts: travis-ci (~travis-ci@ec2-54-163-184-30.compute-1.amazonaws.com) () [00:02:07] *** Joins: travis-ci (~travis-ci@ec2-54-163-184-30.compute-1.amazonaws.com) [00:02:08] (spdk/master) event: allow subsystem to write configuration in async way (Pawel Wodkowski) [00:02:08] Diff URL: https://github.com/spdk/spdk/compare/74f0dfd9b532...2f2dde639400 [00:02:08] *** Parts: travis-ci (~travis-ci@ec2-54-163-184-30.compute-1.amazonaws.com) () [00:03:48] *** Joins: tomzawadzki (tomzawadzk@nat/intel/x-fcdzezpvzxhmnliq) [00:03:55] *** Quits: tomzawadzki (tomzawadzk@nat/intel/x-fcdzezpvzxhmnliq) (Remote host closed the connection) [00:04:01] *** Joins: tomzawadzki (~tomzawadz@134.134.139.75) [00:18:14] *** Joins: travis-ci (~travis-ci@ec2-54-163-184-30.compute-1.amazonaws.com) [00:18:15] (spdk/master) bdev/malloc,null: save uuid when getting config (Daniel Verkamp) [00:18:15] Diff URL: https://github.com/spdk/spdk/compare/2f2dde639400...87f7533a5a5b [00:18:15] *** Parts: travis-ci (~travis-ci@ec2-54-163-184-30.compute-1.amazonaws.com) () [00:24:11] *** Joins: travis-ci (~travis-ci@ec2-54-163-184-30.compute-1.amazonaws.com) [00:24:12] (spdk/master) autotest: update iscsi_filesystem_test (xuhuagen) [00:24:12] Diff URL: https://github.com/spdk/spdk/compare/87f7533a5a5b...da415ae9a025 [00:24:12] *** Parts: travis-ci (~travis-ci@ec2-54-163-184-30.compute-1.amazonaws.com) () [01:00:49] *** Joins: tkulasek (~tkulasek@192.55.54.40) [01:28:45] *** Quits: erodriguez (d8f01e0b@gateway/web/freenode/ip.216.240.30.11) (Ping timeout: 260 seconds) [02:31:39] *** Quits: tkulasek (~tkulasek@192.55.54.40) (Remote host closed the connection) [04:03:22] *** Quits: dlw (~Thunderbi@114.255.44.143) (Ping timeout: 264 seconds) [04:59:51] *** Quits: tomzawadzki (~tomzawadz@134.134.139.75) (Ping timeout: 240 seconds) [05:27:06] *** Joins: tomzawadzki (tomzawadzk@nat/intel/x-acaazspmttumxoyn) [06:12:15] *** Quits: wenyu (67412840@gateway/web/freenode/ip.103.65.40.64) (Ping timeout: 260 seconds) [07:04:13] *** Parts: lhodev (~Adium@66-90-218-190.dyn.grandenetworks.net) () [07:04:16] *** Joins: lhodev (~Adium@66-90-218-190.dyn.grandenetworks.net) [09:25:30] *** Joins: travis-ci (~travis-ci@ec2-54-163-184-30.compute-1.amazonaws.com) [09:25:31] (spdk/master) blobstore: destroy bs_dev on spdk_bs_load fail (Tomasz Zawadzki) [09:25:31] Diff URL: https://github.com/spdk/spdk/compare/da415ae9a025...56cfdb1daa83 [09:25:31] *** Parts: travis-ci (~travis-ci@ec2-54-163-184-30.compute-1.amazonaws.com) () [09:44:29] *** Joins: travis-ci (~travis-ci@ec2-54-224-29-22.compute-1.amazonaws.com) [09:44:30] (spdk/master) rocksdb: remove nonsensical fallocate #ifdef (Daniel Verkamp) [09:44:30] Diff URL: https://github.com/spdk/spdk/compare/56cfdb1daa83...2c7e3a05e3dd [09:44:30] *** Parts: travis-ci (~travis-ci@ec2-54-224-29-22.compute-1.amazonaws.com) () [10:36:57] *** Joins: travis-ci (~travis-ci@ec2-54-160-228-235.compute-1.amazonaws.com) [10:36:58] (spdk/master) bdev/error: Remove config of an error vbdev at the removal of the vbdev (Shuhei Matsumoto) [10:36:58] Diff URL: https://github.com/spdk/spdk/compare/2c7e3a05e3dd...8d57caa11c78 [10:36:58] *** Parts: travis-ci (~travis-ci@ec2-54-160-228-235.compute-1.amazonaws.com) () [10:41:37] *** Quits: alessio (~alessio@140.105.207.235) (Quit: alessio) [10:42:57] *** Quits: tomzawadzki (tomzawadzk@nat/intel/x-acaazspmttumxoyn) (Ping timeout: 240 seconds) [10:44:55] *** Joins: param_ (~param@157.49.169.12) [10:45:17] drv : I have pushed the code. https://review.gerrithub.io/#/c/405489/ [10:45:30] ok, I'll take a look soon - thanks [10:47:01] *** Quits: param_ (~param@157.49.169.12) (Remote host closed the connection) [11:13:46] *** Joins: johnmeneghini (~johnmeneg@pool-100-0-53-181.bstnma.fios.verizon.net) [11:18:09] *** Quits: johnmeneghini (~johnmeneg@pool-100-0-53-181.bstnma.fios.verizon.net) (Ping timeout: 248 seconds) [12:00:08] anyone already have a script for running a variety of sizes, patterns, qdepths for bdevperf that also parses the output into something readable? Easy enough to make one but figured I'd check first... [12:26:26] peluse: looks like there is some perf test stuff in scripts/perf, but only NVMe, not bdevperf currently [12:27:41] drv, thanks. after I get my stuff cleaned up I'll propose it for a bdev subdir in that same area to save anyone else the trouble if they ever want to run a bdev through a perf sweep... [12:27:51] cool [12:28:02] looks like what's there is wired up to use fio with the nvme fio_plugin [12:28:08] and generate CSV output [15:26:34] Hey guys, are there any tips or warnings about running the SPDK in a KVM guest? The first thing/concern in my mind is allocation of hugepages for the guest. [15:29:00] it should generally "just work" - you will need to run scripts/setup.sh as usual inside the guest, and then SPDK programs should work [15:29:34] you can also assign hugepages to back the guest memory, which is a good idea to avoid the host system swapping them out, but that shouldn't be strictly necessary [15:33:23] drv: So, when I do the virt-install for a guest, should I be looking at setting the ram parameter to, say, a minimum of 2GB? [15:34:29] it depends on what you want to run in the guest, but yes, you want to be sure to have enough RAM to assign some to hugepages just like on a physical machine [15:36:32] drv: My plan to experiment/test the SPDK iSCSI target. I'll probably start with it just backed by a ramdisk, but then later use PCI passthru to give the guest an NVMe drive as the iSCSI target. [17:29:56] *** Quits: guerby (~guerby@april/board/guerby) (Ping timeout: 276 seconds) [17:31:51] drv, so wrt scripting some sweeps and capturing/massaging output from bdevperf I think it'll be a lot easier just to change bdevperf to accept some form of scripting instructions (kind like blobcli in script mode) and have an option for CSV output... any thoughts there? [17:33:14] sure, CSV output definitely makes sense [17:33:28] cool [17:33:53] scripting mode sounds like a lot of complexity, though - maybe we can just run a bunch of individual bdevperf runs and concatenate them in bash or whatever? [17:34:03] yeah probably [17:34:19] i was just thinking that after I typed it based on feedback after I did the work in blobcli :) [17:35:05] probably a lot easier actually... [17:36:32] we could do JSON output as well (or instead of CSV), that might be easier to handle if we ever want to extend the information [17:37:07] true, I was thinking CSV as thats easy to graph in excel. could do both too, pretty simple to specify output format and build it either way [17:37:10] I guess we want CSV at the end anyway to import into excel or something like that for graphing, though [17:37:15] hehe [17:37:19] hah, yeah, same idea [18:25:39] *** Joins: dlw (~Thunderbi@114.255.44.143) [18:27:01] *** Joins: travis-ci (~travis-ci@ec2-54-221-15-98.compute-1.amazonaws.com) [18:27:02] (spdk/master) env/dpdk: use MEMPOOL_F_NO_PHYS_CONTIG (Dariusz Stojaczyk) [18:27:02] Diff URL: https://github.com/spdk/spdk/compare/8d57caa11c78...07526d85ef7b [18:27:02] *** Parts: travis-ci (~travis-ci@ec2-54-221-15-98.compute-1.amazonaws.com) () [18:29:01] *** Joins: travis-ci (~travis-ci@ec2-54-160-228-235.compute-1.amazonaws.com) [18:29:02] (spdk/master) doc/virtio: minor updates (Dariusz Stojaczyk) [18:29:02] Diff URL: https://github.com/spdk/spdk/compare/07526d85ef7b...36506448d430 [18:29:02] *** Parts: travis-ci (~travis-ci@ec2-54-160-228-235.compute-1.amazonaws.com) () [21:19:10] *** Joins: guerby (~guerby@april/board/guerby) [22:07:51] *** Joins: weny (67412840@gateway/web/freenode/ip.103.65.40.64) [23:31:22] *** Joins: tomzawadzki (~tomzawadz@192.55.54.44)