January 29, 2019 
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | >
January 29, 2019 [00:01:27] *** newbie56k <newbie56k!6c073c11@gateway/web/freenode/ip.108.7.60.17> has quit IRC (Ping timeout: 256 seconds)[00:04:30] <newbie56k_> Does anyone know how to configure a SAS SSD for active-active dual port? Do I carve up two partitions and access it via each separately, some c0t#sas_address#d0 number?[00:06:10] <newbie56k_> https://documents.westerndigital.com/content/dam/doc-library/en_us/assets/public/western-digital/product/data-center-drives/ultrastar-dc-hc300-series/data-sheet-ultrastar-dc-ss300.pdf[00:06:50] <newbie56k_> See the table Features & Benefits => Benefits column[00:08:27] <rmustacc> newbie56k_: No, the partitions on the device are independent of that.[00:08:43] <rmustacc> Basically it allows you to set up SAS multipathing.[00:08:59] <rmustacc> So, in theory you need two HBAs or expanders that are connected to that in a redundant fashion.[00:12:08] <newbie56k_> interesting, not sure how to hook up two HBAs to a single expander systems. I presume its for doing crazy fast flash SAS JBODs, connecting both path A and path B to the same node.[00:12:31] <newbie56k_> * using dual expander JBODs[00:12:39] <rmustacc> So, an expander may have 8 phys as inputs.[00:12:45] <rmustacc> That you'd normally connect to the HBA.[00:12:49] <rmustacc> They often come in groups of 4s.[00:13:04] <rmustacc> So you'd connect 4 lanes from each HBA to the expander as a bad example.[00:13:17] <rmustacc> But in general, SAS multipathing has a lot of its own challenges.[00:13:35] <newbie56k_> oh yes, i do remember that on the back plane on a few systems. I saw it left un-used.[00:14:24] <newbie56k_> let me verify with the manual[00:17:01] <newbie56k_> ok yea, that is what its for and makes sense given it /is/ a SAS backplane[00:17:03] <newbie56k_> https://www.supermicro.com/manuals/other/BPN-SAS3-216A.pdf[00:17:14] <newbie56k_> page 10, #4 SAS IN ports[00:18:31] <newbie56k_> so if i used a 3x HBA system, to make full use of a active-active dual port performance I need 6x HBA in the system[00:18:33] <newbie56k_> wow[00:23:22] <rmustacc> In general, I'd expect that most folks using an active-active dual port system are doing it for redundancy, not performance.[00:23:26] <newbie56k_> rmustacc: ah yes on the JBODs -- i thought you were talking about the SAS3 back planes[00:24:03] <newbie56k_> I'm not sure how WD would get > 2000 MB/s via only a 12Gb/s SAS path[00:24:35] <newbie56k_> So I think the numbers on the SAS SSD pdf was for pure performance, to boost it up to >= NVMe numbers[00:25:02] <rmustacc> That's fair. They may be doing it to try and counter NVMe in that case.[00:25:12] <rmustacc> But it seems like you have to have two fully redundant paths to that.[00:25:19] <rmustacc> Which is going to be expensive, compared to NVMe.[00:26:02] <newbie56k_> NVMe also has dual-port though I haven't tried it[00:26:37] <rmustacc> Are you intrested in it for redundancy or trying to maximize perf?[00:30:02] <newbie56k_> for NVMe? i'm not sure dual-port will deliver increased bandwidth like the SAS SSDs.[00:30:37] <newbie56k_> for the moment, i'm interested in perf[00:37:25] <newbie56k_> yea, looking at this careful -- you still get more performance/$ out of a box of single port NVMe vs. a box of active-active dual port SAS SSDs using 6x HBAs in a single chassis.[00:40:50] <newbie56k_> so SAS still left around to deal primarily to refresh older gen HA architecture[00:44:28] <rmustacc> If you're trying to maximize perf, sounds kind of what I'd expect.[00:45:34] <Smithx10> rmustacc: do you have a suggestion on how to run NVME workloads for Triton? Would I just have a Pool of NVME on certain CN types?[00:46:32] <Smithx10> in topic box I noticed https://illumos.topicbox.com/groups/zfs/T8fb5137060ef6844-M9303a68ad0408bddbc241364/nvme-jbof[00:48:50] <newbie56k_> Smithx10: yea, I've seen NVMe JBOF from supermicro as well. I haven't eval'd it yet, thought (we)'re still having trouble with NVMe hot swap issues.[00:49:52] <newbie56k_> * in a single chassis system.[00:49:53] <rmustacc> There are a bunch of NVMe hot swap issues still in the OS that we're working on.[00:49:54] <Smithx10> newbie56k_: I haven't ran NVMe in boxes yet, but was just curious on whats the best way to consume flash *[00:50:06] <rmustacc> Smithx10: I think it depends a lot on what you want it for.[00:50:19] <rmustacc> But presumably in a cloud setting, you just want to direct certain workloads there for perf.[00:50:24] <newbie56k_> rmustacc: i just want to say thank you[00:50:35] <rmustacc> No problem, happy to help.[00:52:22] <newbie56k_> Smithx10: so far, (we) use a bunch of SATA SSDs (still cheaper than NVMes) and 2 Intel Optane 375G as ZILs[00:52:41] <Smithx10> Do you trust the sata ssds?[00:53:05] <Smithx10> Over the SAS?[00:53:07] <newbie56k_> yea, we don't use expander (3x HBAs) and datacenter ssds (micron has been good)[00:53:23] <Smithx10> how many disks per hba slot?[00:53:24] <newbie56k_> micron max 5100[00:53:37] <newbie56k_> 8x sata SSDs drives per HBA[00:53:59] <Smithx10> There is not SAS To sata cable?[00:54:03] <Smithx10> no*[00:55:04] <newbie56k_> The supermicro backplane doesn't use SAS-SATA breakout cable though[00:55:14] <Smithx10> gotcha[00:55:17] <Smithx10> What HBA?[00:55:18] <newbie56k_> https://www.supermicro.com/products/system/2U/2028/SSG-2028R-ACR24L.cfm[00:55:20] <newbie56k_> this is the system[00:56:56] <newbie56k_> a newer generation system that includes 4x NVMe[00:56:57] <newbie56k_> https://www.supermicro.com/products/system/2u/2029/SYS-2029U-TRTP.cfm[00:57:18] <newbie56k_> still uses 3x HBAs, and last 4x bays are NVMe[01:21:53] *** Gathis <Gathis!~TheBlack@unaffiliated/gathis> has joined #smartos[01:52:54] *** Gathis <Gathis!~TheBlack@unaffiliated/gathis> has quit IRC (Remote host closed the connection)[01:57:58] *** danmcd <danmcd!~danmcd@208.184.5.170> has quit IRC ()[02:09:20] *** newbie56k_ <newbie56k_!6c073c11@gateway/web/freenode/ip.108.7.60.17> has left #smartos[02:18:23] *** v_a_b <v_a_b!~volker@p57A27D36.dip0.t-ipconnect.de> has quit IRC (Ping timeout: 245 seconds)[02:35:56] *** kkantor <kkantor!~kkantor@c-24-118-59-107.hsd1.mn.comcast.net> has quit IRC (Ping timeout: 240 seconds)[02:42:22] *** richlowe <richlowe!~richlowe@cpe-74-139-197-163.kya.res.rr.com> has quit IRC (Ping timeout: 246 seconds)[02:47:41] *** andy_js <andy_js!~andy@94.3.60.133> has quit IRC (Quit: andy_js)[03:20:55] *** taylorjonl <taylorjonl!c7c0634b@gateway/web/freenode/ip.199.192.99.75> has quit IRC (Ping timeout: 256 seconds)[03:39:54] *** tozhu <tozhu!~tozhu@171.221.44.58> has quit IRC (Quit: tozhu)[03:51:50] *** tuxillo <tuxillo!~antonioh@89.128.97.235> has quit IRC (Ping timeout: 250 seconds)[03:52:07] *** tuxillo <tuxillo!~antonioh@89.128.97.235> has joined #smartos[03:57:24] *** kkantor <kkantor!~kkantor@c-24-118-59-107.hsd1.mn.comcast.net> has joined #smartos[04:02:03] *** kkantor <kkantor!~kkantor@c-24-118-59-107.hsd1.mn.comcast.net> has quit IRC (Ping timeout: 246 seconds)[04:11:58] *** nodens <nodens!~nodens@endor.nodens.org> has quit IRC (Ping timeout: 246 seconds)[04:12:12] *** fiddlerwoaroof <fiddlerwoaroof!~fiddlerwo@unaffiliated/fiddlerwoaroof> has quit IRC (Ping timeout: 246 seconds)[04:19:22] *** jcea <jcea!~Thunderbi@2001:41d0:1:8a82:7670:6e00:7670:6e00> has quit IRC (Quit: jcea)[04:19:52] *** nodens <nodens!~nodens@endor.nodens.org> has joined #smartos[04:21:44] *** fiddlerwoaroof <fiddlerwoaroof!~fiddlerwo@unaffiliated/fiddlerwoaroof> has joined #smartos[04:47:18] *** axonpoet <axonpoet!~axonpoet@fsf/member/axonpoet> has quit IRC (Ping timeout: 250 seconds)[04:51:16] *** yofuh <yofuh!~jw@2a01:4f8:a0:4ffe::2> has quit IRC (Ping timeout: 252 seconds)[05:15:31] *** notmatt <notmatt!uid342873@gateway/web/irccloud.com/x-nkgrmfjlaublbecd> has joined #smartos[05:26:26] *** tozhu <tozhu!~tozhu@101.207.125.86> has joined #smartos[06:58:49] *** lystra <lystra!~lystra@d53-64-11-169.nap.wideopenwest.com> has joined #smartos[08:01:20] *** mno-hime <mno-hime!~mno-hime@94.142.238.232> has quit IRC (Quit: Leaving)[09:15:08] *** v_a_b <v_a_b!~volker@p57A27DE8.dip0.t-ipconnect.de> has joined #smartos[09:16:24] *** notmatt <notmatt!uid342873@gateway/web/irccloud.com/x-nkgrmfjlaublbecd> has quit IRC (Quit: Connection closed for inactivity)[09:25:39] *** kayront_ <kayront_!~kayront@unaffiliated/kayront> has joined #smartos[09:58:33] *** bens1 <bens1!~bens1@161.122.2.81.in-addr.arpa> has joined #smartos[10:14:39] *** man_u <man_u!~manu@manu2.gandi.net> has joined #smartos[11:07:23] *** mojodwar1 <mojodwar1!~mojo@h66-222-27-165.cntcnh.dsl.dynamic.tds.net> has joined #smartos[11:29:21] <sjorge> Anyone want to look at a me change https://cr.joyent.us/#/c/5443/[11:29:33] <sjorge> Simple man fix, no internal jira AFAIK[11:30:47] <jlevon> on it[11:31:55] <jlevon> OS-7545[11:31:57] <jinni> https://smartos.org/bugview/OS-7545[11:34:05] <jlevon> sjorge: got some testing notes for the ticket?[11:45:03] *** andy_js <andy_js!~andy@94.3.60.133> has joined #smartos[11:47:12] *** jrg <jrg!~jrg@unaffiliated/jrg> has quit IRC (Ping timeout: 246 seconds)[11:57:15] *** yofuh <yofuh!~jw@2a01:4f8:a0:4ffe::2> has joined #smartos[11:57:39] *** jrg <jrg!~jrg@unaffiliated/jrg> has joined #smartos[11:58:40] *** tozhu <tozhu!~tozhu@101.207.125.86> has quit IRC (Quit: tozhu)[12:00:31] <sjorge> Doing a build now, will be a few hours[12:00:47] <sjorge> Let me push a change with the correct jira issue[12:05:19] <sjorge> Restarted the build after I pushed the changed commit[12:05:27] <sjorge> Should be able to run 'man vmadm' later tonight[13:26:50] *** tozhu <tozhu!~tozhu@101.207.125.86> has joined #smartos[14:01:28] *** tozhu <tozhu!~tozhu@101.207.125.86> has quit IRC (Quit: tozhu)[14:18:15] <sjorge> rebooting for tests now jlevon[14:18:24] <sjorge> Well once the copy is done :p[14:18:25] <jlevon> ok[14:18:42] <sjorge> test will be: man vmadm, copy example, edit example, apply example to vm[14:23:07] *** sjorge <sjorge!~sjorge@unaffiliated/sjorge> has quit IRC (Quit: 410 Gone)[14:25:31] *** jcea <jcea!~Thunderbi@2001:41d0:1:8a82:7670:6e00:7670:6e00> has joined #smartos[14:29:26] *** sjorge <sjorge!~sjorge@unaffiliated/sjorge> has joined #smartos[14:31:40] *** jcea <jcea!~Thunderbi@2001:41d0:1:8a82:7670:6e00:7670:6e00> has quit IRC (Remote host closed the connection)[14:36:58] <jlevon> sjorge: can you update with Review and approved by lines? I'll push the button[14:41:12] *** jcea <jcea!~Thunderbi@2001:41d0:1:8a82:7670:6e00:7670:6e00> has joined #smartos[15:35:30] *** ruid <ruid!~ruid@unaffiliated/ruid> has joined #smartos[15:51:05] <sjorge> jlevon how do I do that?[15:51:12] <sjorge> I don't think I have every had to do that[15:52:10] <jlevon> https://github.com/joyent/cr-update-reviewers is handy[15:53:04] <rmustacc> You can do it for sjorge if you'd like jlevon[15:53:23] <rmustacc> Otherwise you can edit the commit message yourself with git commit --amend --no-status. Or use a tool like jlevon mentioned.[15:55:56] <sjorge> let me try without the tool first :)[15:59:12] *** kkantor <kkantor!~kkantor@c-24-118-59-107.hsd1.mn.comcast.net> has joined #smartos[16:00:00] <sjorge> OK, done I think[16:05:40] <jlevon> rmustacc: I presume I'm OK to actually press the button even though I CR+IA'd?[16:06:19] <rmustacc> Yeah[16:21:38] *** polishdub <polishdub!~polishdub@207.86.38.254> has joined #smartos[16:35:23] *** kollross <kollross!~kollross@50-82-219-20.client.mchsi.com> has quit IRC (Quit: WeeChat 2.2)[16:45:58] *** gh34 <gh34!~textual@rrcs-162-155-144-114.central.biz.rr.com> has joined #smartos[17:13:30] *** tozhu <tozhu!~tozhu@171.221.44.58> has joined #smartos[17:39:59] *** mno-hime <mno-hime!~mno-hime@94.142.238.232> has joined #smartos[17:42:24] *** mno-hime <mno-hime!~mno-hime@94.142.238.232> has quit IRC (Remote host closed the connection)[17:45:30] *** mno-hime <mno-hime!~mno-hime@94.142.238.232> has joined #smartos[18:07:29] *** Asgaroth_ <Asgaroth_!~Asgaroth@212.2.172.228> has quit IRC (Read error: Connection reset by peer)[18:17:33] *** wizard113_ <wizard113_!~wizard113@astound-64-85-246-199.ca.astound.net> has joined #smartos[18:20:04] <sjorge> thanks jlevon[18:20:08] <sjorge> I think that is a new record :p[18:20:27] <sjorge> fixed around lunch, build done in the afternoon and merged by the evening[18:20:42] <sjorge> To be fair, it was a tiny man page only change :) But shrink to fit.[18:22:01] <wizard113_> question on docker using UDP: I have a container started with '-p 9125:9125 -p 9125:9125/udp' The host that starts the container can hit the UDP port, but another host in the same network cannot. The TCP port works for both.[18:22:12] <wizard113_> Any pointers to what I am missing?[18:22:20] <jlevon> sjorge: my pleasure[18:33:01] *** man_u <man_u!~manu@manu2.gandi.net> has quit IRC (Ping timeout: 244 seconds)[19:22:41] <sjorge> rmustacc: i think you were offline when i posted this earlier[19:23:08] <sjorge> But you seem to do a lot of the acpi stuff[19:23:14] <sjorge> https://gist.github.com/sjorge/d489a6da68e4ee820720938ebcbef225[19:23:26] *** bens1 <bens1!~bens1@161.122.2.81.in-addr.arpa> has quit IRC (Quit: bens1)[19:23:33] <sjorge> Most verbose debug boot output I had ever[19:23:56] <sjorge> Looks harmless but I assume the notices are there for a reason so maybe those are of interest to you, or you know who is[19:25:37] <rmustacc> sjorge: What kind of system is it?[20:12:56] *** andyf <andyf!~andyf@omniosce.ee.ethz.ch> has quit IRC (Remote host closed the connection)[20:12:56] *** hadfl <hadfl!~hadfl@omniosce.ee.ethz.ch> has quit IRC (Remote host closed the connection)[20:16:47] *** andyf <andyf!~andyf@omniosce.ee.ethz.ch> has joined #smartos[20:17:11] *** hadfl <hadfl!~hadfl@omniosce.ee.ethz.ch> has joined #smartos[20:18:03] *** axonpoet <axonpoet!~axonpoet@fsf/member/axonpoet> has joined #smartos[20:21:54] *** black_pete <black_pete!~black_pet@gateway.peterguy.com> has joined #smartos[20:25:23] *** black_pete <black_pete!~black_pet@gateway.peterguy.com> has quit IRC (Read error: Connection reset by peer)[20:36:20] *** mgoetzke76 <mgoetzke76!uid344144@gateway/web/irccloud.com/x-qnkfezogygfxzhfn> has quit IRC (Quit: Connection closed for inactivity)[21:02:50] *** Woodstock <Woodstock!~grumpf@beiss.dessnourst.de> has quit IRC (Quit: ELVIS has left the building!)[21:29:32] <sjorge> Supermicro[21:29:33] <sjorge> Let me[21:29:34] <sjorge> Let me lookup the mobo[21:30:06] <sjorge> SM X9DRi-LN4F+[21:30:17] <sjorge> With a LSI HBA[21:50:01] *** Teknix <Teknix!~pds@2601:801:101:f429:5d3:9286:324e:e7ab> has quit IRC (Ping timeout: 250 seconds)[21:51:10] *** Teknix <Teknix!~pds@2601:801:101:f429:36:1e72:cba:ddd4> has joined #smartos[21:51:25] *** jubal <jubal!~jubal@122-4-237-24.gci.net> has quit IRC (Ping timeout: 244 seconds)[22:03:20] *** richlowe <richlowe!~richlowe@cpe-74-139-197-163.kya.res.rr.com> has joined #smartos[22:30:10] <bahamat> jinni: changelog[22:30:10] <jinni> http://us-east.manta.joyent.com/Joyent_Dev/public/SmartOS/smartos.html[22:34:46] *** Woodstock <Woodstock!~grumpf@beiss.dessnourst.de> has joined #smartos[22:37:56] *** Teknix <Teknix!~pds@2601:801:101:f429:36:1e72:cba:ddd4> has quit IRC (Ping timeout: 240 seconds)[22:38:33] *** Teknix <Teknix!~pds@c-68-50-7-175.hsd1.in.comcast.net> has joined #smartos[22:54:54] *** gh34 <gh34!~textual@rrcs-162-155-144-114.central.biz.rr.com> has quit IRC (Quit: Textual IRC Client: www.textualapp.com)[23:52:58] *** Teknix <Teknix!~pds@c-68-50-7-175.hsd1.in.comcast.net> has quit IRC (Ping timeout: 245 seconds)[23:54:16] *** Teknix <Teknix!~pds@c-68-50-7-175.hsd1.in.comcast.net> has joined #smartos