[00:02] com.farnell.ie would be so... so... american... :-P [01:38] I think it should be com.ie.farnell, but that ship has sailed [02:08] http://pastebin.com/BBgutCEb wtf? [02:08] it's a trap! [02:08] never thought I'd see such a daft error on Linux lol [02:10] heh, "danger will robinson, danger!" [02:10] someone's replaced your 'cd' with a trick one [02:11] don't think so, it works fine before and after that [02:11] oh that's impressive [02:11] or it's failed to write to ~/.bash_history [02:11] I see [02:12] i found my nano_history owned by root today 0o [02:25] type something nice for all the north korean hackers? :-) === zmoylan-1i is now known as zmoylan-pi [04:29] happy birds tweeting outside my window... probably leeching off my wifi... :-) [05:32] ;] [08:40] morning boys and girls. [08:54] 5h37 of deep sleep :D [08:54] allegedly [08:54] 8h of sleep [10:14] Hi all, anyone else having problems logging into graphical mode due to systemd? I'm referring to the bug "starting version 219" [10:15] Bug #1432171 [10:15] bug 1432171 in systemd (Ubuntu) "[udev] Shows "starting version 219" boot message even with "quiet"" [Low,Triaged] https://launchpad.net/bugs/1432171 [10:16] "starting version 219" is the very first thing systemd prints [10:16] just because it is also the last thing you see does not mean systemd is the problem [10:17] Ok, it's not my assessment, it's what cro said in comment #17 [10:18] He says that by going back to upstart everything works fine [10:19] His solutions are: " Either replace the display manager startup scripts to work with `systemd`, or remove `systemd` and replace it with `upstart-sys`" [10:33] Now I can't even see the tty properly [10:39] Can't see anything now, I can still ssh lickily [10:45] Should I just purge xorg and nvidia drivers and try to start from scratch? Is that even the right way to start from scratch short of formatting? [11:27] This graphical problems I'm having are a nightmare [12:05] help, my computer doesnt lock and i want to be the hibernate which i enabled to be remove from the menu [12:21] help, my computer doesnt lock [13:55] anyone else running xubuntu? [13:56] I've noticed that when I go to shutdown my laptop, if I click cancel, it still shuts down. [13:56] o/ [13:57] just clicked cancel... ::waits:: [13:59] one hippopotamus... two hippopotamus... three hippopotamus... [14:00] taps spot where i'd have a wrist watch if i wore one... [14:00] nope, not happening here sorry... [14:01] same for logout, other options? [14:02] StevenR: that's odd, can you report it please [14:02] where exactly did you click on cancel? [14:02] also i assume 15.04 [14:03] trying it in a vm [14:04] doesn't seem to do it here [14:05] touchscreen or mouse/trackpad? [14:06] doesn't do it on the login screen either (it has a different ui) [14:07] do we have an xmpp/jabber client for ubuntu touch? [14:08] ali1234: using the latest LTS [14:09] ok. Can no longer replicate it. hmmph. [14:10] where exactly did you click cancel? [14:10] there's multiple ways to shutdown the computer [14:11] http://catb.org/jargon/html/D/dancing-frog.html :-) [14:11] ali1234: I click the little power logo on the bar, select shutdown, hit cancel, the computer shuts down. At least, that's what happened. Doesn't happen now. [14:12] ali1234: the little logo has other options, like about this computer [14:12] on the bar or on the start menu? [14:12] on the bar [14:12] wait, latest LTS? so 14.04? [14:12] yup [14:13] okay i am looking at the wrong version then [14:13] i don't have a vanilla 14.04 [14:15] * zmoylan-pi just checked... is on 14.04 my kitchen laptop [14:18] i can't see any similar reports [14:25] if one switches channel and one sees... [14:25] ali1234> on the bar or on the start menu? [14:25] on the bar [14:25] ...one does tend to laugh [14:25] xubuntu has a start menu... [14:27] [so do many otther other distros, I was reading "bar" as "pub", much better for the brain] [15:18] Looking at the /topic I see the beer train [RAT] is still prominent... that must be one one heck of a hangover! [15:20] * SuperEngineer is jealous [16:16] https://www.youtube.com/watch?feature=player_detailpage&v=V-pxI8PlY0o#t=1080 [16:16] "proactive suggestions" [16:17] "help the user find value in the store" [16:17] "beautiful pictures" [16:17] no, adverts, adverts, adverts... everywhere [16:26] so buy the hardwarem pay for a commercial os and get stuffed with ads... [16:26] hmmmmm..... how about no [16:30] wow the lag on that stylus sucks [16:31] does anyone know of a normal PC case like this: http://www.scan.co.uk/products/4-bay-25-qnap-ts-451s-nas-solution-with-intel-celeron-dual-core-cpu-1gb-ram-sata-6gb-s-2x-gigabit-la [16:31] those are 2.5" bays [16:31] so cold [16:31] 25c [16:31] ;/ [17:25] 9c :-( [17:28] Hello there, is it possible to get a quick hand configuring my synaptics touchpad with Ubuntu 15.04 [17:29] I would rather appreciate it [17:52] mornin peeps [18:24] ali1234: a PC case with multiple hotswap 2.5" bays? [18:24] which is the same size as that one, yes [18:25] same form factor, but a standard PC that I can install whatever i want on [18:26] er... can't think of anything like that from any of the big players in cases [18:26] and it doesn't have to be hotswap, just front accessible [18:26] nice, 4 x 2.5" in 5.25" space http://www.mypccase.com/icydotomb4x2.html [18:26] yeah, seen those. so then i'd just need a case with 1x 5.25" bay and nothing else [18:27] Not sure you can hotswap in linux though [18:27] you can [18:27] okay, THAT i can find you [18:27] the smallest possible case with a 5.25" external bay? [18:27] i'm interested... [18:28] mATX okay, or you want ITX? [18:29] whichever has the widest selection of motherboards available i guess [18:29] whichever is most "standard" [18:29] mATX by a mile, but mATX is much larger [18:30] mini-ITX is 17x17cm, mATX is up to 24x24cm [18:30] what about NUC? [18:31] nobody's going to put a 3.5" bay on a NUC [18:31] and the motherboard in a nuc isn't something you get control over [18:31] :( [18:31] there are a *limited* number of has-a-cpu-socket mini-ITX boards [18:31] and all micro-ATX boards do [18:32] but let's take http://www.kustompcs.co.uk/acatalog/info_0788.html as an example case [18:32] plus a motherboard from http://www.kustompcs.co.uk/acatalog/ITX_Motherboards.html and the enclosure suggested by OerHeks [18:34] or http://www.kustompcs.co.uk/acatalog/info_0874.html (available in a range of colours) [18:34] or http://www.kustompcs.co.uk/acatalog/info_0849.html [18:34] most mini-ITX cases use laptop slimline optical drives, if any [18:35] i'm going to want something that runs cool and silent too [18:35] how about a pony? [18:35] i don't want any optical drive [18:36] http://www.kustompcs.co.uk/acatalog/info_55245.html - passive cooling [18:36] ali1234: why do you think these cases have any 5.25" bays? just because you're looking to put 4x 2.5" bays in there, doesn't mean that's not why those cases have a 5.25" bay [18:37] because they are designed for HTPCs [18:37] right [18:38] i want to build a silent NAS with SSDs that can also run mythtv [18:38] (backend not frontend) [18:39] okay. do it. i gave you three possible cases, and an array of motherboards [18:42] ali1234: You can get the same CPUs as the NUCs on full size boards [18:43] yeah, that works too [18:43] you'll need to spend the time checking clearance heights for cases vs coolers [18:44] directhex: Well, the passive ones you suggested are good - I've got the older MSI c834 [18:44] oops, c847 [18:44] i am on an MSI kick right now, due to the "Linux out of the box" guarantees [18:45] directhex: Well, in the end it's pretty much an Intel chip slapped on the board - not much else there [18:45] penguin42: ethernet? wifi? vaguely non-broken firmware? [18:46] directhex: Yes (RTL), no, vaguely [18:46] penguin42: i mean in general [18:46] penguin42: if i buy a gigabyte & the firmware has "boot windows & only windows" breakage, their tech support people will tell me to take a flying leap [18:47] directhex: Really it has that kind of breakage? [18:48] penguin42: such breakages exist [18:49] directhex: Yeh, but they are rare and I've found firmware screwups on pretty much everything. TBH the firmware in the MSI doesn't do that much for me, it works but I had some startup problems when I first got it that I'd guess at firmware [18:49] http://mjg59.dreamwidth.org/20187.html [18:49] yeh, I remember that one [18:50] also, NUCs had the same issue in the shipping firmware [18:51] had to update to $latest to boot non-windows [18:53] that does suck, I expect more from Intel, but was it a general non-windows or was it just a screwup? [18:53] that's very different from that Lenovo one [18:53] in 99.9% of cases, it's screwups [18:53] even the lenovo one was a screwup [18:53] we have 40 years of bad BIOS bugs, and the workarounds/fixes to them. we only have a few years of bad UEFI bugs, and the workarounds/fixes to them [18:53] yeh but that was a specific string compare screwup, as opposed to say a device setup that doesn't work with some kernels [18:54] basically, firmware developers are totally terrible at their jobs [18:54] I agree on that [18:54] my main desktop I have to pass a kernel parameter to get modern kernels to boot on it because of some interrupt remapping kernel [18:54] the NUC it was also a string compare error iirc [18:54] maybe i should just buy the QNAP and use their virtualization thing [18:54] i.e. it was hardcoded to only boot Windows Boot Manager from the fixed disk [18:55] directhex: Ah ok, thats dumb [18:56] so it would install, but not boot, linuxes (linuxorum?) in EFI mode. fine with BIOS emulation [18:58] directhex: The EFI boot process isn't exactly simple [18:59] yeah... but the BIOS one is total nonsense in 2015 [19:00] "hello, i am an 8088, i will now boot 16-bit 8088 machine code from the first 448 bytes of the first ATA device on the first ISA bus" [19:00] yeh, technology [19:02] directhex: ATA drive? Hah! [19:02] directhex: It's all actually a lot scarier than that; things like figuring out how many CPUs you've got involves reading values from the RTC [19:02] all insane [19:02] all reasonable bits of evolution at the time [19:03] DOS mode FP exceptions, A20 lines etc etc [19:03] my 6-core 64-bit Haswell-EP is still also an 8088, just in case I want to boot it in the manner of my ancestors [19:04] er, Haswell-E. Haswell-EP is xeon [19:04] yep [19:05] i expect that to go away, with windows 10 basically killing CSM [19:06] directhex: It depends, if they still have to support it working in VMs they might keep the hardware [19:07] penguin42: well Windows 10 logo certified hardware basically won't have it [19:07] directhex: 'wont have it' - you mean it wont have bios compatible boot ? [19:08] yup [19:09] right, that's just the bios module === m0nkey__ is now known as m0nkey_ [19:46] what exactly is the reason that xeons exist? [19:46] ECC support? [19:46] ali1234: ECC, and multisocket on some of them [19:46] ali1234: more cores per package, ECC, Registered, more money per customer? [22:14] ali1234: reasons already given. multi-socket, ECC, more cache, more cores [22:14] my board takes an 18-core chip [22:15] directhex: What board is that? [22:15] penguin42: an X99 board. [22:18] and it lets you take the E5- series Xeons? [22:18] yes [22:20] my old PC would take xeons too. if it can't take a xeon, it's for plebs! [22:20] interesting [22:28] when intel started doing core i3/5/7, i7 was "basically a xeon", with the i7-9xx range. but people moaned about price, and they shipped the "slightly faster i5" series, the "i7-8xx". it's continued to this day - there's "i'm not a xeon, honest" i7's, and "don't look too closely, or you'll see i'm an i5" i7's [22:30] so socket 1366 was used for i7 and uniprocessor xeon; 2011 was used for i7 and uniprocessor xeon; 2011-3 is used for i7 and uniprocessor xeon [22:31] but aren't the 18 processor beasts multisocket only with different busses (CSI?) [22:34] X58, X79 and X99 (the relevant chipsets for the above sockets) use the same bus as xeon. hence supporting xeons :p [22:35] http://en.wikipedia.org/wiki/Intel_X99 [22:38] directhex: What's confusing me is that shows DMI2.0 as the interface from the Xeon, where as I thought Xeons used QuickPath [22:38] * penguin42 is just reading http://www.intel.com/content/dam/www/public/us/en/documents/datasheets/xeon-e5-v3-datasheet-vol-1.pdf [22:40] ah, it has both [22:41] QPI for communication between sockets [22:42] DMI for communication to motherboard chipset [22:42] my old X58 system was QPI for both [22:42] well, except for being uniproc [22:42] but still, QPI for the chipset comms [22:42] directhex: Interesting, I hadn't realised this [22:43] * penguin42 observes from that doc that 'Enable Safe Mode boot' and 'Enable Intel Trusted Execution Technology Agent' are pins on the chip [22:48] they should just make one CPU, it would be so much easier [22:51] ah, the E7's still have a different socket [22:53] hmm, although hmm, [22:54] that data sheet says the E7v2 socket is 'Socket R1' LGA2011-1 [22:54] so that's still the same? [22:54] okay next stupid question: why do they need so many different sockets? they all look pretty much the same [22:54] ali1234: It's Complicated [22:55] ali1234: It sounds like the answer is they're slowly coming to not having many different sockets [22:57] ali1234: but, in essence, it's down to how many physical connections the chip needs to the motherboard - for example, an i7-5930k can drive 40 PCIe lanes, an i7-4790K can drive 16 [22:57] and also you need more pins to get more power into the chip [22:57] the i7-4790k can do dual-channel memory, the i7-5930k can do quad channel [22:58] in server chips, some pins are also for inter-socket links [23:00] IBM POWER8 has 15823 pins (!) [23:01] whaaaaat?!? [23:01] directhex: But your observation, which is interesting, is that now they're just using the same sockets on the xeons as the i7's and just ignoring the inter-cpu links [23:01] directhex: I bet 90% of those are power/ground [23:02] penguin42: not quite 90%, but good guess [23:02] 6000 for power, 7700 for ground [23:02] not far off :-) [23:02] http://www.v3.co.uk/IMG/964/287964/ibm-power8-540x334.jpg?1429512233 [23:02] ali1234: :D [23:03] 2099 for signals, apparently. that's still quite a crazy amount [23:03] they look quite boring compared to the old MCMs [23:04] they also have to be careful just on layout - there's only so many pins you can get out through a PCB from teh centre of a package like that [23:05] well, i would guess that's another reason why so many are power/ground [23:05] penguin42: not if you go deeper! [23:05] since you have to route them on the die as well [23:05] directhex: There'sa limit to how many layers you can sanely do as well [23:05] yep, power and ground will probably get a layer each [23:05] for ali1234 , that's another major issue - a socket with more pins needs more layers, and more layers are MUCH more expensive [23:06] aren't power8 usually used for supercomputers? I think they can probably budget a few extra layers :) [23:06] directhex: Yeh and they start to even get heavy when you get really thick boards - one place I worked did a backplane that was getting silly [23:07] shauno: They're also available as servers - I think 2Uish [23:07] they're billing power8 as a cloud server chip [23:07] directhex: By putting the word 'cloud' on the billing [23:07] :D [23:07] shauno: http://www.tyan.com/campaign/openpower/ [23:08] I can't imagine what makes it particularly cloudy, other than being 8-core [23:09] i'm going to bed [23:09] that's a good place to be