[00:04] hey teward :) [00:05] sarnold: you pinged earlier [00:05] :) [00:05] teward: yeah; a user was having trouble getting nginx to start late enough, ipv6 addresses added via /etc/network/interfaces weren't up yet, so he had to hack up a restart.. [00:05] ah [00:06] sarnold: there's a known 'race' condition there, where if networking isn't up it breaks [00:06] not sure if there's a workaround [00:06] teward: I was curious if you'd seen that, knew a good upstart workaround, or knew if the nginx guys had switched to whatever that magic socket option is that allows binding to addresses that aren't upyet.. [00:06] sarnold: i don't think there's a 'magic socket option' implemented :P [00:07] I *think* there's a way with systemd, but... [00:07] not 100% certain [00:07] that'd be a question sent upstream though [00:07] to Debian or nginx upstream [00:08] teward: aha! [00:08] teward: that poked the neuron that knew that magic socket option: IP_FREEBIND, last paragraph of http://www.freedesktop.org/wiki/Software/systemd/NetworkTarget/ [00:10] sarnold: the tricky part is we ahve to have some kind of network dependency - preferably network-online.target possibly, but not 100%. That'd be a systemctl level command, I believe. And if v4 but not v6 is up, you're still stuck in the "Not Configured" state [00:10] unless systemd has separate v6 networking state commands [00:12] what makes it a little trickier is trying to get that working *earlier* in the boot process [00:12] sarnold: unless upstart has a way to specifically schedule process startup to after all networking is configured? [00:13] teward: heh, good question. :/ [00:14] AIUI there wasn't a way for that in upstart. And AIUI currently, there's no way to differentiate v6 from v4 in 'network is up' state... [00:14] but that's beyond my current knowledge set [00:14] and just my basic understanding [00:15] thanks teward :) === Lcawte is now known as Lcawte|Away === cmagina_ is now known as cmagina === Valfor is now known as Guest43392 === cpaelzer_ is now known as cpaelzer [09:08] What do you use to scan open ports aprt from netstat -tunlp ? [09:10] nmap [09:11] or lsof -i locally === Lcawte|Away is now known as Lcawte [11:37] Hi ... i am facing a Curl issue on my server.. Curl is enabled but not working ...http://awesomescreenshot.com/0445exek37 [11:47] Failed to connect to port 80: Connection timed out.. when using curl [13:56] hello everyone, I was referred to here from ubuntu channel for my question. would using gpt (GUID) partition table on external drives have any drawbacks with existing file systems or OSes? I never used it, and apparently it is preferrable to MBR, but I wish to avoid any unsafe, outright risky situations such as problems recognizing existing data or partitions.. [13:58] caeve: no issues :) [14:00] caeve: The only area you might face problems is for older Windows versions not recognising GPT. I think 64-bit XP is OK, not sure where Vista/7 stand. Windows BIOS installs require MBR, but I *think* they can read GPT for data devices [14:01] TJ- I understand..thank you! [14:04] caeve: If compatibility is your concern, GPT has been supported since Windows XP for both boot and storage devices. [14:05] caeve: Actually, scratch that. Vista and above. [14:08] quantic, there has been mbr in OSes and other devices, so I wanted to be sure things like partition being overwritten because of having been unrecognized or considered free space or similar wouldnt happen [14:09] GPT uses something that things that don't understand GPT will see as a single big MBR partition. [14:09] caeve: Windows XP (potentially) and prior (definitely) will pose that problem. Any Linux/Unix release that's at all recent, and Mac OS X can recognize and read GPT disks. That said, if the partition is formatted with an FS that some other OS doesn't recognize, it may treat it as unformatted and ask to format it itself. [14:09] caeve: as long as you write either a protective or hybrid MBR GPT will be OK [14:09] https://en.wikipedia.org/wiki/GUID_Partition_Table#Legacy_MBR_.28LBA_0.29 [14:10] That page has compatibility information too [14:10] caeve: the *gdisk tools can all do that [14:12] quantic, yes, as it happens every time when ext4 drives are inserted in windows [14:13] TJ- sorry, I couldnt understand the part 'as long as you write either a protective or hybrid MBR' [14:14] do my existing mbr systems need to have these qualities? [14:14] caeve: the GPT tools can write an MBR. Either protective (it has a single partition that covers the entire disk) or hybrid (it has up to 3 partitions - more usually up to 2 - that map onto partitions in the GPT [14:15] caeve: No. The purpose of protective/hybrid MBR when using GPT is to prevent MBR tooling that is not GPT-aware from blindly overwriting/allocating space on the disk that GPT has allocated [14:18] TJ- I am confused as I am not technical enough to understand these terms fully. maybe I should stick with mbr [14:24] hello everybody [14:42] so is it safe to use gpt or not, with my systems being mbr? [14:43] one should not use mbr, unless they are using it for a boot drive on an old bios [14:43] because the extended partition thing os mbr did enough harm [14:43] of mbr* [14:44] patdk-wk, my OS is on an mbr drive, but it is an external drive that I consider using gpt [14:44] on [14:45] if the disk is <2tb, mbr is ok [14:45] if it is >2tb, you should use gpt [14:46] patdk-wk, the question is, is it safe to use gpt with the disk <2tb when the OS on hdd is mbr? [14:46] if you want >4 partitions, gpt could be easier too [14:46] it makes no difference [14:47] patdk-wk, having >4 partitions is the reason why I see gpt as more convenient [14:47] that is like asking, is it safe for me to install water in my house? cause the house down the road has a well [14:47] caeve: For HDDs over 2TB in size, its mandatory to use GPT. [14:47] Capprentice, no, this is not the case for me [14:48] I have used mixed mode. [14:48] so all I want to know is that my OS, which runs on a mbr drive, would not mess any of the partitions or data on the GPT-formatted external drive [14:49] there is no way we can know that [14:49] that will depend on the os [14:49] uncertainty prevails [14:49] but any sane os won't do that, anymore than if you asked it to format your disk [14:49] or if you get a virus [14:49] patdk-wk, it is the latest ubuntu LTS, so I hope it would be reliable? [14:50] yes, it won't touch it, unless told to [14:50] like, you using fdisk/gdisk/sdisk/parted/... [14:51] patdk-wk, ok, I'd like to count on that [14:53] thank you everyone, for your insights and help! === xfailedreality_x is now known as failedreality [16:06] does anyone here know how to check the disks assigned to md0? [16:13] mdadm --detail --scan /dev/md0 (or something along those lines) [16:13] Check the man page ;) [16:18] yeah just got that too. it's very far down the man page [16:18] thanks! === newbie is now known as Guest87222 [17:34] Does this channel cover general ubuntu hosting? [18:08] smb: well never mind, the patch in trusty-proposed is kvm-only, so my qrt runs wouldn't test anything [18:08] or do they? hm [18:09] nope. might be worth augmenting qrt for that [18:11] hallyn, must admit I am not sure right now. might be somewhere in the area of keeping qemu's world in sync with the kvm apic [18:11] i looked at the patch, seems ok [18:12] anyway i'll do a testrun anyway, but i don't see how it could test regressions in this case :) [18:12] but then if i could foresee it, i wouldn't need tests [18:12] very true indeed :) === Piper-Off is now known as Monthrect [19:00] guys, my deluged daemon simply wont run, starting it with $sudo service deluged start seems to work, but status indicates its not running, and web gui shows it offline [19:01] is there a way to get debug log or somethign from the daemon? [19:02] DefunctProcess: check /var/log/upstart/ and /var/log/syslog [19:04] nope === Guest12059 is now known as med_ [19:05] ok [19:06] i think i figured out how to get logs [19:06] now it seems upnp is broken all of a suggen [19:10] smb: huh, actually got a qrt failure. on a nic test [19:10] now i need to revert and see if i can reproduce [19:10] or if it was a fluke === _thumper_ is now known as thumper [19:55] smb: d'oh, yeah, it's an actual regression. holy schnickities === InfoTest1 is now known as InfoTest === Monthrect is now known as Piper-Off [21:29] I am trying to preseed a mini iso with 15.10 downloaded from here -> https://help.ubuntu.com/community/Installation/MinimalCD . This is my preseed file -> http://hastebin.com/jukiyelanu.hs and these are the forms of txt.cfg i have tried -> http://hastebin.com/haquzuzuru.mel So far the mini iso will not use the preseed file for some reason.Does anyone have an idea why? is it not supported with mini iso's? === shauno__ is now known as shauno [22:23] sarnold: ping [22:23] afternoon teward :) [22:23] sarnold: good evening! with regards to the thing a few days ago, upstart's 'network up' check only requires one (either v4 or v6) to be up, I think, which means there's still a race condition [22:23] and I was poking at systemd and came to a similar conclusion [22:24] thanks :) [22:24] so unless Upstart is being revamped to have a v4 and v6 specific option, (same with systemd) that's... blah [22:24] couldn't find anything on IP_FREEBIND upstream though [22:24] so maybe you want to propose that to nginx on trac [22:24] * teward has his hands full with a server migration