=== Lcawte is now known as Lcawte|Away === markthomas is now known as markthomas|away === Laogeodritt| is now known as Laogeodritt === Lcawte|Away is now known as Lcawte [06:43] Good morning. [07:34] Hello guys, I have an Ubuntu server running with a replaced network card. Now I have the problem it doesn't detect the network card, although the hardware is detected by the OS. I've commented out all adapters in /etc/udev/rules.d/70-.... file. Now the adapter appears after 'ifconfig' but it doesn't have an ip. I also set eth0 to dhcp for automatic ip assignment. [07:34] Isn't there a standard procedure to configure a replaced network card? [07:34] Now I don't have an ip address just because I replaced my ethernet adapter :( [07:42] Number5: What is the output of 'ip l'? [07:43] ok, I have to go to the other room, be right back [07:47] well, I get a bunch of line, from lo to eth 3 or 4 Where lo and eth 0 seems to be active and others are in a DOWN state. Could it be that it goes wrong with dhcp? I can try to assign eth0 a static ip. See if that works. [07:47] !pastebin [07:47] For posting multi-line texts into the channel, please use http://paste.ubuntu.com | To post !screenshots use http://imgur.com/ !pastebinit to paste directly from command line | Make sure you give us the URL for your paste - see also the channel topic. [07:48] Wait, I'll check whether the utp cable is plugged in, if that's the case arrggghhhh [07:56] lordievader, I didn't have connection, so that's why I couldn't post the output. [07:58] Right... That does make things harder. Anyhow is the nic you want up? [07:58] Anyway, I solved the problem, because on the backside I found 2 ethernet connections, and one of them was not connected. After I connected the other one the problem is fixed. It's my fault, next time I don't have to rely on my colleague >:( [08:00] He needed the server for his own testing (running windose) after that I don't know what he did. [08:02] lordievader, thnx for trying to help me :) === Locke2002 is now known as Guest48580 [08:27] smb: around? [09:40] hi. I just hit http://askubuntu.com/questions/670106/igb-detected-tx-unit-hang but. I use 14.04 LTS and have spare machine to test [09:41] on production "solved" by downgrading to 3.16.0-46-generic, but it's short term workaround. I wonder how to handle this situation [09:41] will it be fixed in next kernel release? [09:58] Hi guys. Can I do software RAID 10 on Ubuntu ? [09:59] there are some limitation for RAID types ? === orl is now known as mwargh [10:12] mnms_: "Currently, Linux supports LINEAR md devices, RAID0 (striping), RAID1 (mirroring), RAID4, RAID5, RAID6, RAID10 ...", according to mdadm(8) [10:12] mnms_: generally, don't use raid10 - it's a raid level that's somewhat been forgotten about and it lacks the flexibility other levels have [10:12] mnms_: I'd recommend setting up mirrors and using LVM to stripe over those [10:12] RoyK: That is not an unversially agreed upon opinion. [10:12] andol: by whom? [10:13] andol: you can't grow a raid10... which sucks [10:14] RoyK: Which may or may not matter depending on the situation. On the other hand it gives you much better perforamnce than raid5 or raid6. [10:16] andol: you'll get similar performance with lvm striping over mirrors [10:17] andol: but then again, just my opinions here. I really hate RAID sets that can't be changed [10:23] RoyK: raid10 has many advantages. it is advised even for 2 drives [10:24] at least by some people http://blog.a2o.si/2014/09/07/linux-software-raid-why-you-should-always-use-raid-10-instead-of-raid-1/ [10:25] IMHO "cannot grow" isn't really an issue - you also "cannot" grow raid1 or raid5 [10:33] rozie: you can grow *and* shrink raid[56] [10:34] actually, you can grow any raid level except raid10 [10:34] see mdadm(8) [10:34] I would like to create NAS for my backups [10:34] raid[56] is very slow, so out of interest [10:34] and I thought raid 10 will be best option for me [10:35] mnms_: for backup, I'd use raid5 or raid6 - raid[56] is fast enough and raid6 is actually safer than raid10 [10:35] fast enough for backup [10:35] slower on iops [10:35] rozie: depends on the application, it is good for reading [10:35] but fast on squencial i/o [10:36] maswan: it's not bad on writing either - CPUs are fast enough to deal with parity generations these days [10:37] RoyK: As long as it is sequential, random writes or too many streams that it ends up like random is not very fun [10:37] but for reading it is like an n-1 or n-2 stripe in speed [10:38] maswan: well, we have > 200TiB on disk for VM storage, mostly on RAID-6 - it works, and it works well. So please tell me *why* it works ;) [10:39] RoyK: do you have a write-back cache for that raid? [10:39] also, random VMs seldom have high IO demands [10:40] well, had very bad experience with raid5 on hardware controlers. raid10 works like a charm [10:41] mnms_: for NAS - depends how you predict growth. I'd probably just go N times raid1 [10:42] even without lvm, but if you want continous space, use lvm [10:42] rozie: so just simple mirror ntohing more ? [10:42] rozie: I mean without stripe ? [10:43] depends what you NAS will do. but IMO NAS rarely needs performance [10:43] especially write performance [10:44] and raid10 gives you mostly write performance [10:47] rozie: good then, cause mostly I will make a backups and do write operations [10:47] rather then read, occasionaly [10:47] maswan: about 200 VMs currently ;) [10:50] RoyK: 200TB for 200 VPS? it's 1TB/VPS, so probably number of drives does the thing [10:54] rozie: *lots* of different VMs, from tiny to huge. Just trying to illustrate that raid6 works well. You don't need r10 for everything - *most* stuff runs well with raid[56] [10:56] well, have opposite experience. but shared hosting and hardware raid, not VMs and softraid [10:57] rozie: it all depends on your application [10:59] agreed. and raid5 requires checksumming to be done [11:00] but if you have fast CPU and can sacrifice it - may work [11:01] raid5 doesn't require checksumming - it's just parity - and that doesn't consume much cpu, in fact, very little cpu. please don't misinform people [11:02] yep, parity [11:03] I wonder why it was so slow on hardware controllers, then [11:03] (different types) [11:03] because most hardware controllers have a crappy processor and slow/narrow memory buses [11:04] enterprise raid systems always use more or less regular PC systems for their "controllers" [11:10] but returning to my problem: I was hit by kernel bug http://askubuntu.com/questions/670106/igb-detected-tx-unit-hang can I expect it to be fixed in next kernel release? [11:13] RoyK: hardware controllers have gotten better the last 5 years or so, 10 years it was common to get a controller for 12 drives where the maximum throughput of the raid processor in single stream read or write was less than the raw throughput of a single drive. [11:19] maswan: memory bandwidth in PC systems has gotten a wee bit better the last 10 years too ;) [11:19] maswan: so has buses [11:27] RoyK: Yup. I still see limits in the HW raid controllers, but at least it is at aroudn 15-20 disks not 0.5-1 disks in MB/s. Software raid has roughly been faster than spinning disks since pentium III, in my experience. [11:33] RoyK: I hear NVMe SSDs with bcache are all the rage these days [11:37] * patdk-lap wonders how something can be faster than what it's build on [11:38] mirrors you loose half your speed, and other types you loose parity speed [11:38] jpds: just got these https://oc.karlsbakk.net/index.php/s/r7pz44kgJSJmyPG for testing vmware vsan ;) [11:38] patdk-lap: you get better iops with striped mirrors than with parity-based stuff [11:39] you still lost half your iops [11:39] over the raw disk speeds [11:39] patdk-lap: A raid10 of 6 drives should have roughly 3x one drive for write speed and 6x for read speed [11:39] still attempting to understand how softare raid can be faster than the disks it's built on [11:39] patdk-lap: Faster than one individual disk [11:39] maswan: that 6x is highly theoretical ;) [11:40] RoyK: Yeah, I should have said 3-6x. [11:40] maswan, raid should always be faster than a single disk, or you seriously screwed it up [11:40] even back in 386 days [11:40] patdk-lap: Yup. For a long, long time HW raid controllers did. [11:41] even software [11:41] the problem back then, was not the software, but not using a good hba [11:41] patdk-lap: pci-x era raid controllers where a single drive do 100MB/s and the raid controller tops out at 80MB/s... [11:41] the built in ide ports on motherboards are just horrible [11:41] maswan, dunno about your pcix card [11:41] mine got a nice 300MB/sec [11:42] normally maxing out pci-x [11:42] patdk-lap: Then you got lucky. Some of them did go pretty fast, but far from all of them. [11:43] normally used pcix adaptec scsi hba's on my p3's back then [11:43] patdk-lap: plain scsi hbas, or hw raid controllers doing parity? [11:43] RoyK: Not on the openstack bandwagon yet? [11:44] both, normally used hw raid [11:45] jpds: not yet [11:46] Even today, the P822 we got the other day for our VM hosting tops out at roughly 15 of our 25 spinning disks [11:57] coreycb`, hey - I think we'll need to add a dependency on pymysql for all openstack python-XXX's otherwise upgrades will break for existing deployments. [12:55] can someone remind me of the magical command that builds a .deb package automatically from any sources without the debian directory? [12:55] checkinstall [12:55] thanks :D [12:59] rozie: Can you /join #ubuntu-kernel about the igbe bug? [13:04] oh my, third channel... joining right now [13:11] jamespage, ok. I added a pymysql dependency to most of the core openstack packages as well, for sqlalchemy. [13:11] coreycb`, some get it for free via taskflow but better to be explicit I think [13:12] coreycb`, I've jammed most of the dependency bumps that zigo did in experimental into wily-proposed [13:13] coreycb`, that's going to create some instability until we get the b3 milestones uploaded [13:13] jamespage, ok [13:13] coreycb`, oh and I uploaded 2.4.0 of swift yesterday - finally got ontop of the erasurecode dep chain issues [13:13] jamespage, yay! === coreycb` is now known as coreycb [13:16] coreycb, indeed [13:17] coreycb, its an un-optimize erasure coding dep chain, but it is functional [15:10] jamespage, smoser , zul etc: occasionally in an openstack cloud rebooting a Trusty cloud image just hangs at the GRUB prompt. [15:10] no console attached so it is not errant keys [15:10] does this ring any bells? [15:10] (Kilo OpenStack cloud, kvm/libvirt env, Trusty cloud image.) === obsidian is now known as nonparity === utlemmin` is now known as utlemming [15:54] 6/70 [15:59] med_, probably the system didn't fully boot previously and grub is waiting for interaction. [15:59] rbasak, do you recall that bug for grub timeout after failed boot ? === markthomas|away is now known as markthomas [16:04] smoser, thanks. [16:04] smoser: [16:04] https://bugs.launchpad.net/ubuntu/+source/grub2/+bug/1443735 [16:04] Launchpad bug 1443735 in grub2 (Ubuntu Vivid) "recordfail false positive causes headless servers to hang on boot by default" [High,Fix released] [16:04] looking [16:05] it does seem if you're using up to date images you shoudlnt see that [16:07] not sure how up to date ours is but this looks like a strong work around at worst. Thanks Robie and Scott. === kickinz1 is now known as kickinz1\afk [17:26] Hi, i needed to test my servers for ssl2 vulns so i installed openssl packets as in http://www.techstacks.com/howto/enable-sslv2-and-tlsv12-in-openssl-101c-on-ubuntu-1304.html. Now i have problem with skype installation because of no i386 packets. Are there any possibilites to fix this without reinstalling whole system? I'm now on ubuntu 15.04 [17:27] Hello, I'd like to get an ubuntu vagrant box corresponding to the ubuntu server 14.04 installed on my production server (not installed by vagrant). Anyone knows how I could know if I have to get the "-juju" box or the one without "-juju" on this page https://cloud-images.ubuntu.com/vagrant/trusty/current/ ? [17:28] Does anyone know the difference between an install of trusty-server-cloudimg-i386-juju-vagrant-disk1.box and the same *without* -juju in the name? [17:30] lorek123: Are you running 13.04? [17:30] no, 15.04 [17:30] ah. [17:31] lorek123: without answering, I'd strongly suggest using LTS for servers [17:33] marcv: I guess you're faster setting up a server from the "normal" iso :) [17:33] bekks: why do you think so? [17:37] marcv: Because I just deployed a new vm, from the normal ISO, within less than 5 minutes. [17:38] Well, my goal is to get the exact same install (same packages, same versions) as my production server. [17:38] I don't want just any ubuntu [17:39] I only know that my server runs a Ubuntu server 14.04, so my idea is to start the process by installing this exact version [17:40] Install an stock Ubuntu of the same release as your production server, uninstall unnecessary packages, install additional packages matching the software selection of your production server. [17:40] No need for using the very exact same installation medium. [17:40] So just use the stock 14.04 image. [17:42] I think it would be faster to just istall the serve version :-) [17:43] I just had a doubt with this juju issue [17:43] So use the stocl server iso. No need to mess around with juju or vagrant. [17:43] vagrant is a prerequisite [17:43] I want a vagrant box [17:44] but I think I will forget about juju as you say [17:44] You can setup a vagrant box with the stock server iso as well. [17:44] I doubt your production server is a vagrant box, is it? :) [17:46] nope, of course, but there may be no point in generating a vagrant box from the iso if I can get the vmbox directly... [17:46] but I see your point === bdx_ is now known as bdx [19:12] Hi! I can connect to mysql via sequel pro (ssh) but when I do it via my application on the server it doesn't work.. what is a possible solution? [19:12] I got : Access denied for user.... [19:12] Tangurin: Depends on the error message and the configuration of your application. [19:12] So your credentials are incorrect. [19:13] bekks: I use laravel, should I use 127.0.0.1 or localhost? [19:13] Tangurin: laravel means nothing to me. Is the mysql db installed on the same server as your application? [19:13] Tangurin, localhost and 127.0.0.1 should be the same thing [19:13] bekks: yes it is [19:14] SCHAAP137: ok [19:14] I also tried with root and when I write mysql -uroot -p I can write sql queries [19:14] Tangurin: Then you should be able to use 127.0.0.1 or localhost, since both are identical. [19:14] SCHAAP137: there's a slight difference, 'localhost' also resolves to ::1 [19:15] very true, the ipv6 loopback adapter [19:15] forgot about that detail for a moment [19:15] the password is correct I just loged in via sequel pro with the same username and password [19:15] and since ipv6 and ipv4 have unique port ranges, you might have an application that binds to 127.0.0.1:foo but not ::1:foo. [19:16] I can also see the database has the correct name in sequel pro [19:18] somebody? [19:19] Tangurin: does mysql require username / password / source-address tuples in its authentication? [19:20] sarnold: I'm sorry, I don't understand what you mean? I created a user with all possible permissions and with a password. [19:20] Maybe I did wrong in that step? [19:20] Tangurin: But you did not specify a hostname, did you? [19:20] no [19:20] Tangurin: I've never deployed mysql, but postgresql requires setting where that specific usr is allowed to connect from -- I'm curious if mysql does the same, and if it's configured correctly for your application [19:21] Take a look at the users in your db, and you will notice that e.g. root@localhost may login, but root@* may not. [19:21] SELECT User FROM mysql.user; I get one user listed twice? [19:21] Tangurin: See above. [19:21] Tangurin: https://dev.mysql.com/doc/refman/5.1/en/user-account-management.html [19:22] http://laravel.io/bin/yGoK3 [19:23] probably those user names have other fields elsewhere inthe table [19:23] sarnold: Correct. [19:23] .. that mean they are unique when looked at over the entire table [19:23] Tangurin: Nope. [19:23] Tangurin: You are not looking at all relevant fields in the user table. [19:24] Okey I just figured out how bad I know these things, can we go down one level so I understand haha, I don't know how to get the relevant fields [19:24] select * from ... [19:24] Tangurin: I just gave you a link :) [19:24] but I haven't got a clue how mysql stores passwords so don't pastebin the results :) [19:24] bekks: yeah I saw the link thanks, but what am I suppose to look for? [19:25] Tangurin: Well, you have to understand the user concept. [19:25] bekks: I see and I agree but for now because of the deadline I have tomorrow, can you help me without my own knowledge? [19:26] Tangurin: Unfortunately I cant, since I dont know how you setup your users. [19:26] bekks: Can I remove the user and create a new one with correct setup? [19:26] Tangurin: You can, but thats not necessary. [19:26] Read the link, which takes about 15 minutes, and provide all necessary fields for your login. [19:27] Tangurin: I think this might be one of the cases where taking the half-hour or hour to read it through slowly and carefully would actually be the fast route [19:27] bekks: ok. I get a user with the host: "%" what is that? is that giving permission to connect from other than localhost? === adam_g` is now known as adam_g [19:28] sarnold: haha but I am stupid, need more time than that ;) [19:28] Tangurin: hehe :) [19:28] Tangurin: the % is explained here: https://dev.mysql.com/doc/refman/5.0/en/adding-users.html [19:28] Tangurin: we'l, you're here, which is a good start :) if you were _actually_ stupid, you'd be on a web forum instead :) [19:28] sarnold: bekks the annoying thing is that I have done this so many times without a problem but this time I guess I wrote something wrong when I added the user and that might be why it doesn't work [19:29] Tangurin: So read both links carefully, will will take about 20 minutes :) [19:29] sarnold: haha well that was good to hear, now I just have to learn all too! I will, thanks =) [19:30] Tangurin: anyway, there are some times cases where trying to talk with others about the problem is going to be slower than understanding the thing yourself -- and I think this is one of them. there are times when just pasting the error message and getting a quick fix is the easy answer, but database users and application authentication is going to be one of those cases where describing to us what the application needs will take [19:31] sarnold: I understand, and I see by my self that I need more knowledge about this anyway. "give a man a fish and he...." yeah you know that one, I have to hunt my fish so I can have food for a lifetime haha ;) [19:32] Tangurin: there's that, too; plus, once you understand these things, the next time you need it'll come back a lot quicker.. [19:40] sarnold: bekks Sorry I am already back... I just thought about one thing. I use the user "root" now and I can connect via my client application on MAC. And I can see the root as a user when I select all users in ubuntu mysql, how can I possible not login then? I mean I got prove that the username and password is correct and root works for localhost? Isn't it impossible the problem is at the mysql in ubuntu, should [19:40] n't it be in the application which is trying to connect or what? sorry for writing this soon but I just don't understand how it cannot work when it works on my client, and I want to declare this again so you really understand what I am trying to do? [19:41] Tangurin: root has three different hosts, as the user table reveals. [19:41] bekks: yes, correct [19:41] Tangurin: The problem is your login into the mysql db :) [19:41] Tangurin: don't forget that the 'root' of mysql is completely unrelated to the 'root' of ubuntu -- except that connections over the unix socket may require specific unix usernames too... [19:42] sarnold: Yes I understand the difference. bekks so I can login to mysql but not the database. Haha, sorry for beeing this annoying now and for not finish the reading but I get so stresses about this because this deadline is so important and I have never seen this problem before, it has always worked === markthomas is now known as markthomas|away [20:01] So, my NUC (DN2820FYKH) disconnects my USB keyboard somewhere between syslinux and the installer. I'm using trusty's boot.img.gz, any suggestions? :) [20:05] dasjoe: not sure, but have you tried to enable legacy usb in BIOS? [20:07] RoyK: I've reset the BIOS to its default settings, "USB Legacy" is enabled [20:07] dasjoe: not sure, then, sorry [20:07] evening [20:07] Xeth: good localtime(); [20:07] nice 1 :) [20:08] RoyK: thanks anyways, made me re-check whether it's on or off. :) [20:08] i need a bit of help with a VM of ubuntu server, cant seemsto get it right [20:08] !ask | Xeth [20:08] Xeth: Please don't ask to ask a question, simply ask the question (all on ONE line and in the channel, so that others can read and follow it easily). If anyone knows the answer they will most likely reply. :-) See also !patience [20:11] Ok, i have a ubuntu server 15.04 running in a VirtualBox on windows 10, i am using a USB wifi stick for internet and i am not able to get connected to the net on the VM. I did have it working for a brief moment then restarted the server now nothing, have looked at many forums and nothing seems to work. Any ideas? [20:16] Xeth: Are you passing through the USB device to the vm? [20:16] yes [20:18] Is the device recognized in the vm? [20:18] Xeth, any reason you can't use nat on the windows host? [20:20] strange this is, if i do a ifconfig i only see lo and not eth0 or anything else [20:21] Xeth_: Are you using ifconfig or ifconfig -a? [20:23] HELP: How can my ubuntu server krash becuase of an php application without giving me any logs? [20:24] Tangurin: "PHP" :) [20:24] Tangurin: What exactly happened when doing what exactly? [20:25] bekks: I don't want to be lazy but I wrote the problem here so the best is to read it: http://stackoverflow.com/questions/32342654/laravel-kills-ubuntu === markthomas|away is now known as markthomas [20:26] Tangurin: Which other logs besides the apache error log did you investigate? [20:26] bekks: I have tried to fix this for 1-2 weeks now, I really have no idea what the problem can be! If you can help me solve this I will ... I can't say kiss you but I really really need this. This is why I creating a new server every day for trying to see if it works the other day [20:26] Tangurin: TL;DR - which other logs besides the apache error log did you investigate? [20:27] I looked through all los in var/log and also laravel's logs (just a log with php errors) [20:27] Tangurin: And what are the last entries in the messages logfile before the server get killed and the reboot is logged? [20:28] bekks: I have heard about this messages logfile so many times but I have never seen it, what is that? [20:28] Tangurin: /var/log/messages [20:28] bekks: I have none [20:29] So pastebin "ls -lha /var/log" please, as well as "uname -a" and "lsb_release -a" [20:30] bekks: haha you are confusing me... I got som dquote interface when I write that, what shall I do. You are pro! [20:30] bekks: now i see eth0 but still no connection what can i try [20:30] Xeth_: So configure it? [20:31] ok im still new at this, learning how do i do tha [20:31] Tangurin: No. I will not login via ssh and fix your issues. [20:31] that [20:32] bekks: I didn't ask for that, I asked for logging in see the logs because I have watched them so many times [20:32] Tangurin: Use a pastebin then. [20:32] bekks: how do I use the commands you wrote? [20:33] Tangurin: You type in the commands, and copy and paste the output to a pastebin site, then provide the URL to us. [20:34] bekks: http://laravel.io/bin/W4D3Y [20:38] Tangurin: Can you pastebin the dmesg.1.gz file? [20:38] bekks: it is not readable? just question marks and strange characters [20:39] gz is compressed [20:39] how do I copy the content+ [20:40] !info gist-paste [20:40] Package gist-paste does not exist in vivid [20:40] oops [20:40] !info gist [20:40] gist (source: gist): Upload gists to gist.github.com. In component universe, is optional. Version 4.3.0-1 (vivid), package size 13 kB, installed size 87 kB [20:40] !info pastebinit [20:40] pastebinit (source: pastebinit): command-line pastebin client. In component main, is optional. Version 1.4-4 (vivid), package size 14 kB, installed size 164 kB [20:41] two packages that provide that functionality [20:42] zcat dmesg.1.gz | pastebinit [20:43] thanks, I guess I did it correct? [20:43] bekks: http://paste.ubuntu.com/12316438/ [20:44] Tangurin: Try to pastebin the kern.log please [20:45] bekks: how do I write if it is not a compressed file? [20:47] bekks: http://paste.ubuntu.com/12316455/ [20:49] Hi guys, how do you guys recommend to remote your ubuntu server from a windows machine? [20:50] arcsky: USe Putty, and ssh in :) [20:55] bekks: This problem only occured at this hosts server but not when I use digitalocean server and I install ubuntu and its packages the same, so I don't understand how it only works at digitalocean and not at the host I use here === wmp is now known as Guest70264 === Lausi_ is now known as Lausi [21:10] bekks: i mean GUI [21:10] arcsky: You mean GUI in which context? [21:15] bekks: I'm imaginind zentyal or similar [21:17] bekks: like RDP in windows [21:17] arcsky: Use freenx then. [21:17] ist better then vnc? [21:18] Everything is better than VNC. === Guest70264 is now known as wiuempe === wiuempe is now known as Guest99445 [21:34] also depends on what VNC [21:34] it's extendable, so in theory somebody could fix it? :) [21:35] and there is SPICE, of course [21:36] JanC: VNC is insecure and slow by design. [21:36] it's not insecure over a secure connection [21:37] most wrappers offer ssh tunneling for vnc [21:37] VNC is insecure - you just tunnel it through a secure tunnel. [21:37] bekks: yes. http is also insecure, you just tunnel it thru ssl [21:38] correct. [21:38] and AFAIK there already are several extensions that make it somewhat better, but I don't know _how_ good it can get if somebody puts enough time in it === Guest99445 is now known as wmp [21:40] also most VNC issues are related to 3D desktops & SFX, right? === Lcawte is now known as Lcawte|Away === Tm_Tr is now known as Guest8229