[02:28:17] *** Joins: tkulasek (tkulasek@nat/intel/x-rxbldlvrftuonqoz) [03:22:47] *** Quits: tkulasek (tkulasek@nat/intel/x-rxbldlvrftuonqoz) (Remote host closed the connection) [05:40:25] *** Joins: darsto_ (~darsto@89-78-174-111.dynamic.chello.pl) [05:41:27] *** Quits: darsto (~darsto@89-78-174-111.dynamic.chello.pl) (Ping timeout: 240 seconds) [05:41:27] *** darsto_ is now known as darsto [06:10:55] *** Joins: darsto_ (~darsto@89-78-174-111.dynamic.chello.pl) [06:12:15] *** Quits: darsto (~darsto@89-78-174-111.dynamic.chello.pl) (Ping timeout: 252 seconds) [06:12:15] *** darsto_ is now known as darsto [06:15:57] *** Joins: darsto_ (~darsto@89-78-174-111.dynamic.chello.pl) [06:16:47] *** Quits: darsto (~darsto@89-78-174-111.dynamic.chello.pl) (Ping timeout: 240 seconds) [06:16:47] *** darsto_ is now known as darsto [07:17:28] *** Joins: johnmeneghini (~johnmeneg@pool-108-20-29-249.bstnma.fios.verizon.net) [07:39:25] *** Joins: darsto_ (~darsto@89-78-174-111.dynamic.chello.pl) [07:40:33] *** Quits: darsto (~darsto@89-78-174-111.dynamic.chello.pl) (Ping timeout: 250 seconds) [07:40:33] *** darsto_ is now known as darsto [07:44:17] *** Quits: ppelplin2 (uid326872@gateway/web/irccloud.com/x-nltutputbxsnwvtc) (Quit: Connection closed for inactivity) [07:47:14] *** Quits: johnmeneghini (~johnmeneg@pool-108-20-29-249.bstnma.fios.verizon.net) (Quit: Leaving.) [07:49:35] lhodev: will you modify https://review.gerrithub.io/#/c/spdk/spdk/+/429963/ to install needed apps? [07:50:50] Yes, I'll do that [07:57:35] ok [07:58:28] and for python issue: it reply to in trello on the motivation about switching to python3 [07:59:31] but for you case, I think we can create separate package or use conditional build in RPM spec file. [08:00:34] eg "--with default-python" or something [08:03:18] then we can sed shebang during rpmbuild [08:03:30] how about that lhodev? [08:12:18] *** Joins: johnmeneghini (~johnmeneg@pool-108-20-29-249.bstnma.fios.verizon.net) [08:12:47] *** Quits: johnmeneghini (~johnmeneg@pool-108-20-29-249.bstnma.fios.verizon.net) (Client Quit) [08:13:19] jimharris: about your comments in https://review.gerrithub.io/#/c/spdk/spdk/+/429910/ [08:13:54] I think you should rather look at final implementation here https://review.gerrithub.io/#/c/spdk/spdk/+/430095/ [08:14:52] I can add mutitreaded tests if you wish, but would like to leave that spinlock there. [08:15:21] * ... mutitreaded tests ... :P [08:19:39] with this patch serie reading JSON configuration seem to work. I'm working on tests but I have not completed them yet. Will have them tomorrow. [08:25:57] pwodkowx: I read your responses regarding python (https://trello.com/c/mrasr8mc/9-spdk-tools). While your proposal of modifying the spec file with a conditional to employ just "python" would get me past my rpmbuild failures on my distro, I wish we could come up with some other way of dealing with the shebang's than using sed to alter those source python scripts that were changed explicitly to python3. [08:28:27] I might have a way of using the python scripts "as-is" -- i.e. with the ref's to python3 -- in another manner. The key issue for me is to eliminate the "Requires: python3", or as you stated, do it conditionally such that it can place the dependency on "python" instead of "python3". [08:29:23] I have to run to a meeting. Will check back with you afterwards, possibly via Trello or something since I know your business day is quickly approaching a close. [08:32:24] pwodkowx - looking... [08:40:22] pwodkowx: i still don't understand how this locking handles the case where two different threads are calling spdk_jsonrpc_client_poll() simultaneously [09:54:36] *** Joins: pwodkowx_ (540a1326@gateway/web/freenode/ip.84.10.19.38) [09:59:13] *** Quits: pwodkowx_ (540a1326@gateway/web/freenode/ip.84.10.19.38) (Ping timeout: 256 seconds) [10:30:23] *** Joins: travis-ci (~travis-ci@ec2-54-144-80-50.compute-1.amazonaws.com) [10:30:24] (spdk/master) bdev/rbd: fix init error message (Patrick Ohly) [10:30:24] Diff URL: https://github.com/spdk/spdk/compare/2d718da0438c...5b02ea4767ac [10:30:24] *** Parts: travis-ci (~travis-ci@ec2-54-144-80-50.compute-1.amazonaws.com) () [10:33:02] *** Joins: travis-ci (~travis-ci@ec2-50-17-65-213.compute-1.amazonaws.com) [10:33:03] (spdk/master) bdev/rbd: fix error handling after failed rados_connect (Patrick Ohly) [10:33:03] Diff URL: https://github.com/spdk/spdk/compare/43b335f9177b...40da9ab5e360 [10:33:03] *** Parts: travis-ci (~travis-ci@ec2-50-17-65-213.compute-1.amazonaws.com) () [11:23:47] Recd notification that Piotr set me as the Assignee for https://review.gerrithub.io/#/c/spdk/spdk/+/430538 Looking at Gerrit docs, it's unclear to me precisely what that means. Does that imply that for any additional changes I would like to propose (or anyone else), that I essentially take over this change, perform the edits and submit the patch? [11:38:39] jimharris: ping [11:40:33] hi lhodev: assignee just means that we're requesting you to drive the review for this patch [11:40:59] which you seem to already be doing :) [11:41:07] Thanks. I hope I didn't offend Piotr with so many of my rephrasing suggestions ;-) [11:41:23] oh no, he won't be offended [11:42:26] Unrelated, I'm in a quandary over this whole python/python3 mess. [11:43:01] *** Joins: darsto_ (~darsto@89-78-174-111.dynamic.chello.pl) [11:44:15] *** Quits: darsto (~darsto@89-78-174-111.dynamic.chello.pl) (Ping timeout: 252 seconds) [11:44:15] *** darsto_ is now known as darsto [12:02:34] *** Joins: darsto_ (~darsto@89-78-174-111.dynamic.chello.pl) [12:03:38] *** Quits: darsto (~darsto@89-78-174-111.dynamic.chello.pl) (Ping timeout: 246 seconds) [12:03:39] *** darsto_ is now known as darsto [12:43:10] lhodev: I definitely won't be offended :-) [12:43:36] all feedback appreciated [12:44:12] Thanks ppelplin [12:45:19] do you think I should add anything else to this patch/section about shared build? [12:45:38] or we can push the patch in current form? [13:05:23] ppelplin: I just gave it a +1 [13:06:05] jimharris: bwalker: Can one of you guys take a quick look at: https://review.gerrithub.io/#/c/spdk/spdk.github.io/+/429984/ [13:17:29] crap - I think our filter just looks for "RFC" anywhere in the commit message (or at least anywhere in the first line of the commit message) [13:24:19] i wanted to see how your patch rendered [13:33:26] jimharris: are you going to massage the filter? [13:48:36] i'm going to have sethhowe massage the filter :) [13:59:24] hey guys, just looking at jenkins [13:59:43] I see that last few patches failed because of same error [13:59:49] /usr/local/qemu/spdk-2.12/bin/qemu-system-x86_64: error while loading shared libraries: libspice-server.so.1: cannot open shared object file: No such file or directory [14:02:13] Pawel K and me are looking into it, but we're not sure how to fix it yet [14:03:13] lhodev: It's not because of the RFC in the commit message (at least on the commit message) I am parsing for [RFC]. There is something else going wrong. Can you please rebase and upload another revision of the patch? [14:03:31] We'll probably disable vhost_autotest until we fix it [14:04:32] sethhowe: ok [14:09:09] sethhowe: Dang, I had already amended that commit locally while removing the [RFC] part in preparation for submitting it to CI and review. [14:10:27] sethhowe: I guess I could amend again adding the '[RFC]' line if you wanted to test out that filter logic. [14:11:10] lhodev: You don't need to change the commit message or anything, you can just pull latest master and rebase your patch on it. [14:11:26] That will be enough to allow you to push another rev [14:12:46] sethhowe: I think you're missing the point. I had *already* earlier amended that patch (Change 429963) locally with a couple other changes and elimination of 'RFC' in the title in preparation for submitting for actual review per Pawel's request. [14:14:30] Wait, I thought the issue is that even though the patch doesn't have [RFC] in the title, it still isn't running through the test pool. [14:15:47] In the build pool logic, we search for the string '[RFC]' (not just 'RFC'). So I don't think the contents of your commit message are what is keeping it from running through the test pool. [14:16:25] How does the commit message for patch set 2 compare to the commit message you have locally? [14:16:27] i'm an idiot - this is a spdk.github.io patch [14:17:19] sorry for the thrash [14:18:18] haha gotcha. Yeah, I can confirm that it's running here. https://ci.spdk.io/spdk.github.io/status/ [14:18:44] lhodev: jimharris: I'll add lhodev to the whitelist for the spdk.github.io pool. [14:19:31] rendering looked good - I just merged it (it takes a day or so to sync out to spdk.io) [14:19:39] thanks [14:19:49] no - thank you! [14:22:55] jimharris: pwodkowx: Just pushed https://review.gerrithub.io/#/c/spdk/spdk/+/429963 this time with the [RFC] removed. It also adds the string magic that converts the example name based on the directory components relative to SPDK_ROOT_DIR. [14:24:26] jimharris: Note there is a couple ugly inconsistencies. examples/bdev/hello_world produces a binary during the build of "hello_bdev" -- not "hello_world". Similarly, examples/blob/hello_world produces "hello_blob" [14:24:59] Now, I didn't add an "install:" target for the aforementioned so they're unaffected. [14:26:24] However, if one were to instrument them with the new INSTALL_EXAMPLE macro, then they would be renamed according to the directory paths; e.g. "examples/bdev/hello_world/hello_bdev" would install as $DESTDIR/usr/local/bin/spdk_bdev_hello_world [14:29:43] As discussed yesterday, I added the install targets just for two examples: nvme/perf and nvme/identify [14:33:36] oh - you made that very fancy :) [14:35:54] Lol, I'll take "fancy" for $200 Alex. I 'spose others might look at that and slap their foreheads thinking, "what the....." :) [14:36:04] what if for now we just added INSTALL_EXAMPLE_APPNAME? and nvme/perf and nvme/identify could specify those as nvme_perf and nvme_identify respectively [14:36:41] then for 19.01 we could rip off the band-aid and change the actual app name to nvme_perf [14:36:48] just an idea [14:36:54] what you have there is fine for now [14:37:17] We could, but with the programmatic version I put in place, it reduces the chance that someone might accidentally give INSTALL_EXAMPLE_APPNAME a name already in use :) [15:27:14] *** Joins: peluse (~peluse@134.134.139.72) [15:27:14] *** ChanServ sets mode: +o peluse [15:27:42] bwalker, thanks for updating my reset patch!! I like the tweaks however... [15:27:58] I think you addressed 2 unrelated items in that patch ;) [15:31:58] which things [15:33:24] nothing huge (a) movement of the driver init function, unless this was related to resets but I guessed it was more about QAT in CI (b) freeing of mempools that was previously missing. Again, thanks for the updates though, really appreciate it, has been a tough week to say the least [15:33:36] oh - that's because I rebased it [15:33:39] I didn't actually change that [15:33:54] * peluse should have known better :) [15:34:04] that's a fix we merged yesterday from Seth to deal with a DPDK bug [15:34:20] no one tests multiprocess in DPDK I guess [15:34:22] yup yup, I'm a bit behind on history for the last few days.... [15:34:29] certainly no one has loaded QAT in a secondary process at least [15:34:35] but we sure do! [15:34:39] wrt mp, yeah I know I think that's awesome - way more real-life for testing [15:35:07] we mostly do it to improve test times so we don't keep initializing hardware [15:35:10] but it finds a ton of bugs [15:35:25] yeah, I also had that one liner DPDK bug I pushed that was for the same reason, nobody had ever used any vdev I guess in a secondary (or at least not w/any kind of memory leak detection) [15:36:41] anyway, happy to cover for you. There's a zillion patches in the queue, but once it goes through and passes it's the last one in and we'll do the 18.10 release [15:37:43] I actually installed a QAT card to test and all of the tooling to set that up made it really trivial [15:37:52] it's all very slick [15:37:58] and it worked flawlessly [15:42:43] Excellent!!! [16:08:06] whats going on with fedora-03? [16:08:27] it's gone offline a bunch of times today [16:08:35] sethhowe has been fighting with it [16:09:36] who is winning? [16:09:53] fedora-03 [16:09:57] blow out [16:10:26] heh [16:10:40] I can't figure out what's wrong with it either though [17:12:17] *** Joins: Shuhei (caf6fc61@gateway/web/freenode/ip.202.246.252.97) [17:14:12] jimharris: There are two issues I should work on, fio_plugin with DIF and raid bdev. The latter may be caused by my patch series. The latter looks higher priority than the former. I'll work on the raid bdev first. [18:12:29] jimharris: I couldn't reproduce the issue of the raid bdev. I'll wait for repu1sion's update. [20:03:51] *** Quits: Shuhei (caf6fc61@gateway/web/freenode/ip.202.246.252.97) (Ping timeout: 256 seconds)