| 00:27.03 | *** join/#utah amjith (~amjith@c-24-10-150-142.hsd1.ut.comcast.net) |
| 01:57.24 | *** part/#utah bbeattie (~bbeattie@208.53.57.89) |
| 02:17.45 | *** join/#utah igotnolegs (~igotnoleg@75-162-83-166.slkc.qwest.net) |
| 02:17.55 | *** join/#utah TimRiker (~TimRiker@bzflag/projectlead/TimRiker) |
| 02:17.55 | *** mode/#utah [+o TimRiker] by ChanServ |
| 02:33.47 | *** join/#utah tiwula (~lane@63-248-12-52.static.sdyl0420.digis.net) |
| 02:40.37 | *** join/#utah nisovin_ (~nisovin@64.55.129.66) |
| 02:47.15 | *** join/#utah mabes (~mabes@68.69.166.132) |
| 02:51.26 | *** join/#utah amjith (~amjith@c-24-10-150-142.hsd1.ut.comcast.net) |
| 03:45.20 | *** join/#utah fugue88 (~dsowen@173-165-137-65-utah.hfc.comcastbusiness.net) |
| 04:38.54 | *** join/#utah ScytheBlade1 (~Death@serenity.averageurl.com) |
| 04:38.54 | *** join/#utah ScytheBlade1 (~Death@about/pxe/ScytheBlade1) |
| 05:17.19 | *** join/#utah gerberb (~gerberb@opensuse/member/Gerberb) |
| 05:52.53 | *** part/#utah TimRiker (~TimRiker@bzflag/projectlead/TimRiker) |
| 08:22.03 | *** join/#utah BLZbubba (~mark@tpsit.com) |
| 11:51.34 | *** join/#utah maquis (~maquis@c-67-168-18-46.hsd1.wa.comcast.net) |
| 11:51.44 | *** join/#utah niso (~nisovin@64.55.129.66) |
| 11:52.37 | *** mode/#utah [+v maquis] by ChanServ |
| 12:01.43 | *** join/#utah maquis (~maquis@c-67-168-18-46.hsd1.wa.comcast.net) |
| 12:01.43 | *** mode/#utah [+v maquis] by ChanServ |
| 14:01.39 | *** join/#utah emcnabb (~emcnabb@cpe-065-190-176-153.nc.res.rr.com) |
| 14:01.40 | *** mode/#utah [+v emcnabb] by ChanServ |
| 15:17.03 | *** join/#utah brkpt (~brkpt@204.238.46.101) |
| 15:19.35 | *** join/#utah macnewbold (~macnewbol@76.8.223.101) |
| 15:24.44 | *** join/#utah tvc123 (~tcameron@host-148.pl1071220.fiber.net) |
| 15:28.47 | *** join/#utah BLZbubba (~mark@tpsit.com) |
| 15:29.38 | *** join/#utah fadein (~Erik@c-67-161-246-186.hsd1.ut.comcast.net) |
| 15:50.55 | *** join/#utah jalbretsen (~jalbretse@pandorica.itransact.com) |
| 15:58.10 | *** join/#utah TimRiker (~TimRiker@bzflag/projectlead/TimRiker) |
| 15:58.10 | *** mode/#utah [+o TimRiker] by ChanServ |
| 16:04.30 | *** join/#utah neybar (~jalance@74-220-196-116.bluehost.com) |
| 16:05.15 | *** join/#utah fungus (~olsonl@bromine.sosstaffing.com) |
| 16:07.06 | *** join/#utah mahelious (~mahelious@ip65-46-59-10.z59-46-65.customer.algx.net) |
| 16:15.41 | *** join/#utah craigj (~craig@neutroninteractive-xmission.xmission.com) |
| 16:18.42 | *** join/#utah pdayton (~Adium@199.188.28.9) |
| 16:18.47 | *** join/#utah mabes (~mabes@24-104-67-234-ip-static.hfc.comcastbusiness.net) |
| 16:28.50 | *** join/#utah mrpull (~Jake@216.58.143.230) |
| 16:32.05 | *** join/#utah mabes_ (~mabes@24-104-67-234-ip-static.hfc.comcastbusiness.net) |
| 16:38.50 | *** join/#utah pelaofeliz (~pelaofeli@67.137.236.66) |
| 16:39.11 | *** join/#utah tristanbob (~quassel@2001:1948:212:0:f890:e75c:6896:a9c6) |
| 16:51.47 | *** join/#utah saturdayplace (~thunsaker@nat.spillman.com) |
| 16:52.59 | saturdayplace | <PROTECTED> |
| 16:53.10 | saturdayplace | Cat pictures. Work every time. |
| 16:53.26 | jima | ... |
| 16:53.40 | brkpt | That, and fart jokes |
| 16:54.27 | *** join/#utah sjansen (~sjansen@ip65-44-231-178.z231-44-65.customer.algx.net) |
| 16:57.29 | *** join/#utah sjansen (~sjansen@ip65-44-231-178.z231-44-65.customer.algx.net) |
| 17:09.19 | *** join/#utah mabes (~mabes@24-104-67-234-ip-static.hfc.comcastbusiness.net) |
| 17:09.52 | *** join/#utah nibb13 (~nibb13@206.71.87.254) |
| 17:32.48 | *** part/#utah TimRiker (~TimRiker@bzflag/projectlead/TimRiker) |
| 17:33.09 | eightyeight | has anyone successfully configured gnu/linux nfs4 tcp to get > 100 MBps writes? |
| 17:33.29 | eightyeight | gigabit lan, of course |
| 17:34.10 | sjansen | gnu/linux? really? |
| 17:34.15 | sjansen | shuns eightyeight |
| 17:36.06 | eightyeight | loves pushing that button. you always get an entertaining response out of someone |
| 17:41.52 | fozzmoo | sjansen++ |
| 17:59.39 | *** join/#utah zachwily (~zach@ec2-184-72-231-143.compute-1.amazonaws.com) |
| 18:03.59 | *** join/#utah pdayton1 (~Adium@199.188.28.9) |
| 18:08.20 | fugue88 | eightyeight: I read your comment as > 100 mega/mibi bytes per second (you used a B, not a b). |
| 18:08.44 | fugue88 | In which case, you're pretty close to the max for a 1 giga bit per second link, no? |
| 18:08.51 | eightyeight | b = bits, B = bytes |
| 18:09.03 | fugue88 | Exactly. |
| 18:09.17 | eightyeight | no. currently, i'm at 70 MBps |
| 18:09.42 | fugue88 | In any case, with framing overhead, I wouldn't expect anyone to get > 100 MBps. |
| 18:10.08 | fugue88 | Or did you mean "approaching 100 MBps" instead of "greater than 100 MBps"? |
| 18:10.16 | eightyeight | $ wget http://server.on.lan/4G-file.iso -O /dev/null |
| 18:10.22 | eightyeight | yields 105 MBps |
| 18:11.13 | eightyeight | iperf(1) also shows a solid, full, saturated 1 Gbps pipe |
| 18:11.29 | fugue88 | Okay, so we're arguing about 5MBps. Great. |
| 18:11.41 | eightyeight | i want 30 |
| 18:12.05 | eightyeight | my point is, my nfs mount is not saturating the pipe |
| 18:12.20 | eightyeight | 100, 105, 125, whatever. i'm no where near it |
| 18:12.52 | eightyeight | thus, my question: has anyone succeded in this area, and if so, how? |
| 18:14.55 | Sargun_Screen | What |
| 18:15.01 | Sargun_Screen | What are you trying to achieve? |
| 18:15.09 | jima | "moar bits" |
| 18:15.14 | eightyeight | 100 MBps sustained throughput |
| 18:15.27 | eightyeight | to disk, ignoring write cache |
| 18:23.34 | eightyeight | what tools can i use to monitor network latencies as through put increases for an nfs mount? |
| 18:23.38 | eightyeight | iostat(1)? |
| 18:24.37 | eightyeight | i need to look more closely at iops, rather than raw throughput |
| 18:27.06 | jima | did you enable jumbograms? |
| 18:27.21 | goozbach | jumbo frames! |
| 18:27.32 | goozbach | (sung to the tune of chocolate rain) |
| 18:28.20 | eightyeight | aren't jumbo frames really only useful to prevent data fragmentation? |
| 18:28.24 | eightyeight | s/data/packet/ |
| 18:28.39 | eightyeight | doesn't know a lot about jumbo frames |
| 18:29.10 | jima | i thought they somewhat could improve throughput |
| 18:29.18 | jima | alas, i don't, either :-P |
| 18:29.51 | eightyeight | ~ping JoshH |
| 18:29.52 | ibot | pong JoshH |
| 18:44.23 | eightyeight | is now enlightened about jumbo frames after visiting JoshH's office |
| 18:50.55 | eightyeight | digs through /proc |
| 18:52.14 | goozbach | "Are jumbo frames enabled?" is my company's "Have you turned it off and back on?" |
| 18:52.38 | Sargun_Screen | SJFs, good sir.. |
| 18:53.42 | Sargun_Screen | eightyeight: 1) Stop using NFS (use AFS) |
| 18:54.00 | eightyeight | why? |
| 18:54.07 | JoshH | Sargun_Screen: go back to begging for an MX80 |
| 18:54.18 | Sargun_Screen | JoshH: :( |
| 18:54.30 | Sargun_Screen | eightyeight: It's a protocol that's outlived its usefulness. |
| 18:54.47 | Sargun_Screen | JoshH: The systems guys just spend $150k on 20TB of storage, and I can't get an MX80. |
| 18:54.55 | JoshH | terrible |
| 18:55.22 | fugue88 | Sargun_Screen: If AFS, why not Coda? |
| 18:55.27 | JoshH | just have the storage guys include one on their next unnecessary order |
| 18:56.11 | eightyeight | i would do infinband, if application support didn't suck, and hardwire drivers weren't so buggy and scarce |
| 18:56.27 | eightyeight | s/wire/ware/ |
| 18:57.22 | eightyeight | iscsi may be our next option |
| 18:57.41 | goozbach | AoE? |
| 18:57.49 | goozbach | :) |
| 18:58.00 | eightyeight | *shrug* |
| 18:58.22 | fugue88 | goozbach: Age of Empires? That might be a good alternative to all this budget non-sense... |
| 18:58.34 | fugue88 | Ah, ATA over Ether. |
| 19:02.43 | sjansen | goozbach is obviously biased, but AoE just makes sense |
| 19:02.55 | goozbach | yes and yes |
| 19:03.13 | sjansen | iSCSI feels like the J2EE of storage to me |
| 19:03.34 | goozbach | sjansen++ |
| 19:03.35 | fugue88 | Well, *that* permantently tainted it for me. |
| 19:06.53 | fugue88 | Sargun_Screen: Still interested in your opinion about Coda. Also, do you use the in-kernel AFS? |
| 19:10.23 | *** join/#utah tiwula (~lane@70.89.246.241-Busname-UT.hfc.comcastbusiness.net) |
| 19:22.43 | *** join/#utah pdayton (~Adium@199.188.28.9) |
| 19:26.24 | Sargun_Screen | So, let's take a few steps back - why do you need a network based filesystem |
| 19:26.51 | eightyeight | shared storage for a virtualization infrastructure |
| 19:26.59 | Sargun_Screen | Oh. |
| 19:27.10 | Sargun_Screen | What's your backend |
| 19:27.41 | eightyeight | 10 3 Gbps 7200 RPM 64 MB cache 2TB SATA drives in a manual software RAID 1+0 |
| 19:27.57 | eightyeight | iozone(1) reports 700 MBps random writes |
| 19:28.23 | eightyeight | that is one storage server |
| 19:28.42 | eightyeight | another will be built exactly like it, and replicated over 10 Gig-E via DRBD |
| 19:29.00 | goozbach | eightyeight: which virt software? |
| 19:29.23 | eightyeight | each storage server has 4 gigabit ports, in a 802.3ad bond |
| 19:29.23 | fungus | for virt shared storage, NFS can work, but not stock implementations. Stock Linux or Solaris NFS servers can't cut higher I/O reqs of virtualization. IMHO |
| 19:29.35 | eightyeight | goozbach: libvirt+qemu |
| 19:29.41 | eightyeight | (kvm) |
| 19:29.43 | goozbach | running on? |
| 19:29.49 | eightyeight | ubuntu 10.04.3 |
| 19:29.53 | goozbach | rhel/cenos debian/ubuntu |
| 19:30.02 | eightyeight | server-edition |
| 19:30.24 | fungus | I'd suggest iSCSI, pretty easy to setup, free, and decent. |
| 19:30.31 | sjansen | Quit wasting your time with that Ubuntu crap and switch to FreeBSD already. ;-) |
| 19:31.09 | goozbach | how much total storage are you needing/using? |
| 19:31.31 | eightyeight | 10 TB in that RAID setup |
| 19:32.02 | goozbach | anticipated future expansion? |
| 19:32.17 | sjansen | fungus: What sort of iSCSI? |
| 19:32.33 | eightyeight | currently 4 hypers with dual bonded gigabit nics to the switch. may expand a couple more |
| 19:32.38 | Sargun_Screen | eightyeight: you're doing it wrong |
| 19:32.42 | sjansen | 'Cause there's an embarrassment of options. |
| 19:32.57 | Sargun_Screen | eightyeight: So, what 10G platform are you using? |
| 19:33.03 | eightyeight | targeted io is random read/writes with small data sets- mostly sql inserts/selects |
| 19:33.36 | fungus | sjansen, unknown, i've only played w/ open-iscsi |
| 19:33.42 | eightyeight | Sargun_Screen: intel |
| 19:33.44 | goozbach | I mean total cap of storage |
| 19:33.53 | eightyeight | goozbach: unknown |
| 19:33.56 | Sargun_Screen | eightyeight: So, for virtualization, AFS and Coda haven't been too good to me, they end up eating up way too much CPU, and seem to have surprisingly high latency when doing lots of block mainpulation |
| 19:34.00 | Sargun_Screen | eightyeight: I meant, what switch? |
| 19:34.06 | Sargun_Screen | Does it support DCB-CEE? |
| 19:34.17 | eightyeight | Sargun_Screen: cisco something. JoshH could tell you more |
| 19:34.37 | goozbach | do you have a backup infrastructure already in place? |
| 19:35.03 | eightyeight | this is a test infrastructure i'm building |
| 19:35.13 | eightyeight | i can tear down / build up as i please |
| 19:35.42 | Sargun_Screen | eightyeight: He works for you? |
| 19:35.48 | eightyeight | s/for/with/ |
| 19:35.57 | Sargun_Screen | JoshH: What are you hooking eightyeight up with? |
| 19:37.30 | goozbach | this would get you 16-48T with up to 6-1G aggreate bandwidth |
| 19:37.32 | goozbach | http://www.coraid.com/products/srx2800_3.5_sas_sata_etherdrive_storage_appliance |
| 19:37.40 | goozbach | or optional 10G infrastructure |
| 19:38.09 | JoshH | Sargun_Screen: it's a couple stacked 3750X's |
| 19:38.22 | Sargun_Screen | JoshH: Why not QFX? |
| 19:38.25 | JoshH | just to connect the kvm boxes to the storage rig |
| 19:38.39 | goozbach | <PROTECTED> |
| 19:38.58 | Sargun_Screen | eightyeight: So, I would setup Gluster with a replica set of 2, and utilize RDMA, and then mount local gluster client as NFS |
| 19:39.11 | Sargun_Screen | ubut, I don't know if the 3750X supports RDMa |
| 19:44.28 | Sargun_Screen | but, that'll get you low-latency, failover, high throughput, sexiness. |
| 19:45.08 | Sargun_Screen | goozbach: How would he share virtual block devices? |
| 19:46.46 | goozbach | <PROTECTED> |
| 19:46.51 | goozbach | one of two ways |
| 19:47.29 | goozbach | divide the shelf into multiple luns, and using masks seperate them |
| 19:47.55 | goozbach | use some sort of clustering FS and share all the storage with all hosts |
| 19:48.48 | Sargun_Screen | Can the appliance manipulate luns hot? |
| 19:48.54 | Sargun_Screen | Like, if I want to grow/ shrink them |
| 19:48.56 | eightyeight | i think iscsi is next |
| 19:49.22 | Sargun_Screen | eightyeight: Opinion on my proposal? |
| 19:49.55 | Sargun_Screen | I mean, I don't know your I/O per VM requirements, but I could suggest some other changes |
| 19:50.17 | eightyeight | i don't have a lot of time to re-architect everythnig from scratch |
| 19:50.26 | goozbach | Sargun_Screen: not this appliance, but you can do so if you throw a VSX into the mix |
| 19:50.38 | Sargun_Screen | goozbach: So, I'm curious as to how the VSX works |
| 19:50.53 | eightyeight | i was hoping to have a working solution by tomorrow with this setup |
| 19:52.20 | sjansen | I'm not a network storage expert, but my first question would be "is the throughput you're hoping for even possible with NFS?" |
| 19:52.59 | eightyeight | i'm guessing not. i thought it might, even with some /proc/sys/ tuning, but it appears i'm hitting the ceiling |
| 19:53.24 | eightyeight | i've not needed this sort of nfs performance before |
| 19:54.00 | eightyeight | aft |
| 19:54.06 | sjansen | Putting database storage on NFS just seems wrong to me. |
| 19:56.09 | jima | yeah, i think i'd be more prone to use iscsi/aoe |
| 19:56.45 | jima | well, assuming the db in question allows raw block access |
| 19:57.15 | jima | i'm not sure how much merit there is in throwing a filesystem into the mix |
| 19:58.00 | Sargun_Screen | sjansen: So, NFS can actually do pretty well when you tune it |
| 19:58.13 | Sargun_Screen | but, for his problem case, NFS doesn't seem to make sense. |
| 19:58.30 | Sargun_Screen | goozbach: is the VSX just CLVM with a fancy badge on it? |
| 19:59.25 | sjansen | I think most of their stuff os Plan9 based, so it's probably not CLVM exactly. |
| 20:01.39 | Sargun_Screen | o.O |
| 20:01.46 | Sargun_Screen | I thought plan9 was dead. |
| 20:02.09 | goozbach | Sargun_Screen: not CLVM |
| 20:02.17 | sjansen | No, Plan 9 is from outer space. |
| 20:02.46 | goozbach | it's more of a cross between LVM and NAT applied to ATAoE |
| 20:03.01 | fungus | eightyeight, for your demo purposes and proof of concept. iSCSI is simple and will work. You can investigate new options later. |
| 20:03.13 | goozbach | it utilizes the fact that an AOE reponse doesn't need to come from the target it was sent to |
| 20:03.59 | JoshH | it was iscsi a long time ago before eightyeight got here, and it did work fine. there's just been a lot of 'rethinking' of it since then |
| 20:06.42 | *** join/#utah kendsnyder (~Adium@67-2-145-168.slkc.qwest.net) |
| 20:08.23 | eightyeight | there's more than db storage on this storage, for sure |
| 20:08.37 | eightyeight | it is the target though |
| 20:40.21 | *** join/#utah bgerber (~gerberb@suse103.zenez.com) |
| 20:51.10 | goozbach | TIL about ss |
| 20:51.17 | goozbach | which apparently is the successor to netstat |
| 20:51.20 | goozbach | hmm |
| 20:52.30 | JoshH | whoa |
| 20:53.05 | eightyeight | hmm |
| 20:53.13 | eightyeight | iscsi only allows one connection per lun |
| 20:53.23 | eightyeight | how do i enable shared storage? |
| 21:05.00 | *** join/#utah brkpt (~brkpt@wbucrp-gdm0a-as.bsc.disney.com) |
| 21:05.03 | Sargun_Screen | eightyeight: Expose the same lun over multiple iscsi luns |
| 21:05.06 | Sargun_Screen | underlying lun |
| 21:06.06 | *** join/#utah mrpull (~Jake@216.58.143.230) |
| 21:07.01 | eightyeight | Sargun_Screen: have some docs i can read on that? |
| 21:18.39 | fadein | okay ninjas, I've got a C code question for ya: |
| 21:19.07 | fadein | Is there a standard C function that will translate "/home/fadein/./derp.txt" into "/home/fadein/derp.txt" |
| 21:19.17 | fadein | a la Java's File.getCanonicalPath() |
| 21:19.30 | sjansen | yes |
| 21:19.58 | fadein | sjansen: do you know its name? |
| 21:20.43 | sjansen | Trying to. |
| 21:20.50 | sjansen | remember |
| 21:21.29 | sjansen | man realpath |
| 21:23.05 | fadein | sjansen++ |
| 21:23.13 | fadein | you have saved the ranch! |
| 21:43.08 | fadein | lame - realpath() only works if the supplied filename is an existing file |
| 21:43.12 | Sargun_Screen | eightyeight: uhh, sory, don't. |
| 21:43.58 | eightyeight | no offense, but i didn't think so |
| 21:44.26 | Sargun_Screen | I wasn't sure if you were being sarcastic, or serious |
| 21:44.56 | eightyeight | i've gotten a lot of "do this" and "do that", without really anything to back it up |
| 21:45.43 | Sargun_Screen | eightyeight: the iSCSI thing is just pretty much common knowledge |
| 21:46.08 | eightyeight | then finding docs on nested luns should be easy. ツ |
| 21:46.43 | Sargun_Screen | I think I probably found it on our storage server's manual |
| 21:46.44 | Sargun_Screen | (EMC) |
| 21:46.54 | Sargun_Screen | The glusterfs thing was from an internal engineering architecture that we created. Unfortunately, it's under some wraps |
| 22:00.18 | *** join/#utah kendsnyder (~Adium@67-2-145-168.slkc.qwest.net) |
| 22:01.34 | levi | Silly sjansen, that's not a standard C function. It's a standard POSIX function. |
| 22:06.22 | levi | Someone in scrollback mentioned using AFS instead of NFS. Is there some new AFS that came out recently? |
| 22:06.30 | levi | CIFS all the way, man. ;P |
| 22:07.38 | levi | Huh, I guess someone stuck Andrew File System in the linux kernel. Go figure. |
| 22:10.35 | fugue88 | Marked experimental, and the docs on the config option disagree with the docs in Documentation/filesystems/afs.txt. |
| 22:11.07 | fugue88 | At least so on 3.0.6. |
| 22:12.30 | fugue88 | Same deal in 3.1.6 looks like. |
| 22:22.14 | levi | It's crazy that we're still using AFS and NFS. |
| 22:22.44 | eightyeight | how do you setup AoE for shared storage? |
| 22:23.02 | eightyeight | it appears that you don't want to mount the target from more than one client, or fs corruption could occur |
| 22:31.15 | sjansen | levi: Yeah, but I suspect anyone asking the question doesn't care about C vs POSIX. |
| 22:33.39 | levi | You never know. A lot of my software development team has Windows boxes, and we use mingw to run unit tests on stuff that also gets cross-compiled to QNX, and some others use Linux. There are a lot of things you'd assume would be compatible across the three OSes that are not. |
| 22:34.07 | sjansen | eightyeight: Again, not a stroage engineer but as I understand it that's the role of a clustered FS like Gluster or GFS. |
| 22:35.53 | eightyeight | or clvm |
| 22:35.57 | eightyeight | yeah. that's what i'm finding |
| 22:36.03 | sjansen | levi: I don't expect that type of developer to ask for "a standard C function" since they'd be fully aware that C itself has very little to say about paths. |
| 22:38.06 | eightyeight | device -> raid -> clvm -> aoe |
| 22:38.18 | fadein | levi: I don't care about STD C vs POSIX |
| 22:39.05 | fadein | I care about "is this function likely to come along with GCC" |
| 22:39.11 | shalkie | CLVM doesn't protect against FS corruption. It merely propogates LVM metadata. |
| 22:39.27 | shalkie | You still have to use a FS that understands clustering eightyeight |
| 22:40.30 | eightyeight | but, software raid doesn't understand cluestered filesystems, that i can see |
| 22:41.41 | sjansen | Seems to me you're trying to do RAID at the wrong layer. |
| 22:41.56 | eightyeight | i'm trying to understand the layers |
| 22:42.03 | shalkie | sjansen: Thinking like me that it should be done with the LVM device? |
| 22:42.18 | eightyeight | my head hurts |
| 22:42.19 | shalkie | hasn't investigated that option... |
| 22:42.42 | Jayce^ | http://www.youtube.com/watch?v=SuHmEo0Bx7Q&feature=youtu.be |
| 22:42.43 | sjansen | I would expect client -> aoe -> host -> raid |
| 22:42.53 | Jayce^ | this had BETTER be real |
| 22:43.03 | sjansen | suddenly realizes eightyeight is probably trying to achieve multipathing |
| 22:43.20 | eightyeight | sjansen: "host" meaning the storage server, or the hypervisor? |
| 22:43.35 | eightyeight | i want multiple hypers accessing one storage pool |
| 22:43.36 | sjansen | storage server |
| 22:44.38 | sjansen | Jayce^: That had better be a movie and not a marketing campaign. |
| 22:44.47 | eightyeight | so, then client{1..4} -> gfs -> aoe -> host -> raid |
| 22:46.11 | sjansen | yeah |
| 22:46.46 | eightyeight | and s/aoe/iscsi/ would work as well, no? |
| 22:47.33 | sjansen | With the caveat that aoe is limited to a broadcast domain and iscsi isn't, even though taking advantage of that is probably a bad idea. |
| 22:50.29 | sjansen | Jayce^: I can already feel my nerd rage growing. No way it can be good. |
| 22:50.53 | shalkie | Jayce^: I expect to see a new Kit Kat commercial coming up soon now... |
| 22:52.35 | Jayce^ | I really want to see the three of them skipping work.. |
| 23:06.15 | *** join/#utah wps (~wps@wfw.vivint.com) |
| 23:16.35 | shalkie | Jayce^: I am sorry, I was thinking of Twix and not Kit Kat. |
| 23:45.12 | fadein | off to work! |
| 23:46.46 | fadein | er, that is, off to school |
| 23:46.52 | fadein | been a long day, can you tell? |