[00:00] jordanrandles: are there any errors or warnings logged in any log files that look relevant? [00:02] sarnold: I wouldnt know, I dont have access to the log files. The server computer is in my server room but it wont let me input anything using a keyboard as if the USBs are not working but they are fine. Its ever since I 'tried' setting up openvpn [00:12] Its not taking any input from the keyboard [00:23] it seems like the server only does this using Linux 3.16 [00:24] But I just booted using Linux 3.13 and it works fine. Is there any way to make the server always boot into this version? [00:26] you could edit your grub menu to only include the 3.13 kernel, or uninstall the 3.16 kernels.. [00:27] but more likely that's just a symptom of something else that didn't work, perhaps the initrd images weren't properly created for the 3.16 kernels.. [00:28] So what do I do to fix this? [00:28] Or find out if that is the problem? === markthomas is now known as markthomas|away [02:21] No internet from my server need help [02:21] I had to go earlier and was not able to fix the issue [02:23] jvwjgames: start with the basics: make sure ethernet wires are fully seated, not wrapped around flourescant ballasts, etc.; make sure everything in the path is properly plugged in; check dmesg and log files for errors [02:23] jvwjgames: make sure you've got good link light from both NICs; replace endpoints if you need to [02:24] jvwjgames: check ip addr show and ip route show output on both hosts, make sure they make sense [02:24] Whenever I ping or traceroute I get interface IP as reply regardless of what the destination is [02:24] jvwjgames: start with simple protocols, ping to IPs, then nc to tcp and udp ports; if those work but larger packets don't work, maybe pmtu settings are broken? [02:25] See my post above ^^ [02:27] jvwjgames: is the IP address assigned by DHCP, or manually set? [02:27] Manually [02:28] And I do know it's right cause I tested it before [02:28] jvwjgames: does "ip addr show" and "ip route show" look correct on both hosts? [02:28] jvwjgames: can you pastebin (transfer via USB maybe) the "/etc/network/interfaces" file if that is where you've defined the interface [02:29] Ok === _Jeepbeats is now known as Jeepbeats [03:43] I'm having issues with a domain user authenticating when I try to log on [03:43] the workstation has been joined to the domain [03:49] hi guys, does anyone know where UFW gets started in the boot process? where can I find out? === CiPi is now known as cipi === cipi is now known as CiPi [07:04] i'm having problems enabling htaccess on my apache2 ubuntu 14.04 [07:06] i have default install and have done .htaccess and .htpasswd with a webgenerator site, it says "Forbidden" currently the dir with .htaccesss [07:07] i have AllowOverride All [07:07] dont have htpasswd program that many guides refer to [07:10] now have htpasswd from apache2-utils [07:14] did htpasswd -c /var/www/html/poly/.htpasswd poly [07:15] so everything should be okay [07:16] Good morning [07:17] morning lordievader [07:18] -> cig [07:18] Hey neurotus [07:19] lordievader: any help with apache2 .htaccess would be appreciated === wmp is now known as Guest59283 === Guest59283 is now known as wiuempe [07:21] What are you trying to accomplish? [07:22] lordievader: basic user-auth [07:24] Ok, where is the directory defined? [07:25] That is where I have configured the basic auth. [07:26] /var/www/html/poly [07:28] neurotus: That is the directory, not apache's definition of it. [07:28] Those are usually somewhere in /etc/apache2/ [07:39] darn, now the whole /var/www/html is protected and doesnt ask for password and not only the html/poly folder [07:41] How did you configure the basic auth? [07:45] i get 403 FORBIDDEN and it doest not ask for a password [07:46] How did you configure the basic auth? [07:46] lordievader: i just put .htaccess in /var/www/html/directory and put in apache2.conf AllowOverride All [07:46] So you didn't configure anything? [07:46] and did .htpasswd in directory [07:46] if you remove .htaccess does the folder work [07:51] Ben64: nope [07:51] then fix that first [07:54] morning all [07:59] -> away === jordan_ is now known as T3DDY === trick is now known as Guest81882 [10:36] i get "Forbidden" when i try to access my apache2, have purged apache2-data and apache2 and done a clean install [10:37] i'm gonna purge them again and move /var/www as it isn't empty and check /etc [10:50] ok, now working, had to do something with my contents of /var/www/html [10:52] as they were not removed on purge of apache2 [11:10] what was that package that checked after apt-get update that no processess are using old files ? [11:12] coreycb, I re-merged python-pbr from experimental but I had todo it the hard way [11:12] downloaded the source package directly from ftp.debian.org and did it by hand [11:12] coreycb, I think we can do 'syncs' the same way === Lcawte|Away is now known as Lcawte [12:27] coreycb, syncs might be functional again in a bit - launchpad team are working on it [13:08] does anyone here use openvpn-as? [13:09] jamespage, ok good thanks [13:09] does anyone know if there is a way to get openvpn-as on a intel server?? [13:15] | [13:21] coreycb, ok syncs are possible again now - ceilometerclient done :-) [13:21] jamespage, \o/ [13:22] coreycb, I think you can sync all of the clients yourself right? [13:22] they are mostly/all seeded [13:22] coreycb, ping me if you hit something not right [13:23] jamespage, ok going to take a pass through them now [13:23] coreycb, awesome [13:23] that should make the core packages for rc1 build and uploadable [13:24] does anyone know if there is a way to get openvpn-as on an intel server?? [13:26] openvpn-as? [13:27] Openvpn-access server. All I can find are packages for AMD servers but I have intel servers [13:28] jordan_: why don't you just use the oss openvpn? [13:28] jordan_: amd64 is the architecture intel uses these days [13:28] ;-; I dont know how to setup the normal openvpn package [13:28] jordan_: there are lots of docs out there, and it's not hard, also openvpn is free === Lcawte is now known as Lcawte|Away [13:28] openvpn-as is not [13:29] I see, Ill try and set it up because of you [13:29] :) [13:32] fails with 'dh1024.pem': No such file or directory ????? [13:36] jordan_: what does? [13:42] jordan_: Did you create that file? [13:42] jordan_: It's part of the openvpn RSA set up process [13:49] i have 8 gb ram with 3 guest and 2 guest with memory ballooning but it seems it is not using bollooning [13:49] it is using swap file [13:50] on kvm 14.04.3 [14:01] coreycb, problem with pysaml2? [14:02] jamespage, yeah, haven't gotten back to it today but for some reason tests seemed to be running code in the .pc/ dir [14:03] samba35: how much memory have you given those guests? [14:03] 64kb [14:03] 2 gb and 2gb allocated with 1 gb ballon [14:04] coreycb, is that required for keystoneclient? [14:04] jamespage, yes [14:05] 64kb is suppose to be use ? [14:05] patdk-wk: ? [14:05] coreycb, are we sure? [14:05] jamespage, checking [14:05] coreycb, i can't see it in either requirements or test-requirements [14:05] coreycb, zigo had this wrong before [14:06] in the packaging that is [14:06] jamespage, ok yeah let me drop it and try building [14:06] coreycb, zigo also added memcached - I lack context for why that's been added [14:07] well, if your using cp/m or an apple][, sure [14:07] samba35: he's probably just trolling as usual ;) [14:08] :) [14:08] jamespage, they both get hits when grepping the code [14:10] coreycb, really? [14:10] * jamespage looks again [14:10] jamespage, I think memcache is optional for tests [14:11] coreycb, I really cant see any evidence of requirement for those two deps in keystoneclient [14:11] keystone - yes - client -no [14:11] jamespage, I don't think saml2 is needed, the only import of saml2 happens to be a keystoneclient file named saml2 [14:12] yes [14:12] jamespage, I'm dropping them and rebuilding [14:12] ok [14:14] smb, is the dpdk in the NEW queue exactly the same as in your PPA? [14:15] jamespage, It should be, though I forgot whether arges has not made some minimal changes like removing comments about fixme [14:16] smb, I really need that into the archive [14:16] I have todo the ovs bits as well [14:16] and I'm running out of time window [14:16] jamespage, we need some aa or are you one? [14:24] is there a chance to do a a 2.22 -> 2.23 bump for python-novaclient https://bugs.launchpad.net/cloud-archive/+bug/1499510? [14:24] Launchpad bug 1499510 in python-novaclient (Ubuntu) "In Kilo Upgrade python-novaclient to 2.23" [Undecided,New] [14:24] or should I just do a request for the specific fix I need? [14:25] btw, do I want to get it into vivid first, and then it get's synced to CloudArchive? [14:35] jamespage, ok so its not identical to the ppa but to what is in the lp git tree (ubuntu-wily branch) [14:36] jamespage, beside the fixme there is one relevant change to call dh_install in binary indep which causes the examples to be packaged [14:53] Would anyone know why a md raid-10 array have this type of io pattern? http://paste2.org/YDLGaLJW [14:57] mnaser: what's in /proc/mdstat? [14:58] RoyK: http://paste2.org/La4aHdHm [14:58] mnaser: no idea - perhaps there's only a little space used and all that is on the first four drives? [14:59] (first two mirrors= [14:59] (first two mirrors) [15:00] RoyK: it's a fresh new server so yes, but shouldn't reads come from all stripes? [15:00] mnaser: perhaps better to ask on #linux-raid [15:00] mnaser: I've never used r10 with md [15:01] RoyK: thanks, i'll have a poke there === markthomas|away is now known as markthomas [15:58] coreycb, as I tagged keystone, I uploaded just now [15:59] jamespage, thanks [15:59] coreycb, heat versioning can be fixed up and re-uploaded as we got the broken one dropped from proposed [15:59] jamespage, ok [15:59] jamespage, keystoneclient py3 tests are not happy, 40 failures [16:00] jamespage, most of the other clients are in good shape now though [16:00] coreycb, ack === Lcawte|Away is now known as Lcawte [16:09] coreycb, that should unblock alot of core package uploads - test everything via PPA first and then shove them in the queue for release team to review [16:13] jamespage, ok. I'm testing in here btw: https://launchpad.net/~corey.bryant/+archive/ubuntu/wily [16:14] coreycb, awesome === CiPi is now known as cipi === cipi is now known as CiPi === Lcawte is now known as Lcawte|Away [18:40] hello, I moved the content of / of my mdadm raid lvm volume to a new hard drive partition and installed grub manually on the new hard drive, the system is running but I cant start several services and when I try to use tab for autocompletion in directory navigation it tells me: -bash: cannot create temp file for here-document: Permission denied [18:41] what is wrong here? The syslog file is also clear :\ [18:42] nch: feels like you might have incorrect permissions on a directory somewhere [18:42] nch: /tmp feels quite likely; try ls -ld /tmp [18:42] sarnold: it gives me that: "drwxr-xr-t 5 1777 root 4096 Sep 25 20:40 tmp" [18:43] aha, that's probably it ;) [18:43] nch: chown 0:0 /tmp ; chmod 1777 /tmp ; check it again [18:45] now it gives me that: drwxrwxrwt 5 root root 4096 Sep 25 20:45 tmp [18:46] but it is still not working [18:46] sarnold: any other idea why syslog gets no entries and my apps are not starting due to permission errors I guess? [18:47] nch: hmm, you may have more permissions problems; I was hoping that'd be the only one [18:48] I don't know off-hand any way to automate looking for them on ubuntu system [18:48] sanold: but why? I just copied the whole content of / with cp :\ === wmp is now known as Guest36842 [18:49] nch: cp doesn't copy permissions and owner and group information unless you request it specifically with e.g. cp -a or fiddle with --preserve ... [18:49] well shit -.- [18:49] nch: if you copied a whole system withuot cp -a you may find it easier to start over === bitbyte_ is now known as bitbyte [18:50] Ok now I dont have the original / is there any way to change the permissions for the whole system? [18:50] nch: another possibility is to fix errors one at a time until you get a point tha tyou can run apt-get install --reinstall ... on every package on the system [18:52] sarnold: well shit [18:52] nch: that sounds like a pretty fair summary of the situation :) [18:53] sarnold: well yes, I have no plan what to do now :\ [18:53] nch: I'm actually surprised it booted far enough for you to get to a shell that lets you solve problems. it could be worse. [18:54] nch: but there's no denyuing that fixing this will be painful. === TheEternalAbyss is now known as Guest44026 [18:54] sarnold: yes the boot was no problem, also samba and its network shares are working fine but the rest not^^ [18:55] nch: check ls -ld output for a huge number of directories; /etc /etc/init* /etc/rc* /bin /usr/bin/ /sbin/ /usr/sbin/ etc etc.. [18:56] nch@media:/$ ls -ld /etc drwxr-xr-x 112 root root 4096 Sep 25 20:03 /etc [18:56] nch@media:/$ ls -ld /bin drwxr-xr-x 2 root root 4096 Sep 25 04:45 /bin [18:56] sarnold: well that looks good so far [20:14] I’m trying to setup a raid using mdadm ( i don’t have dmraid installed) and i have created 4 partitions using fdisk. When i try to create the array i get ….. [20:14] openstack@openstack:~$ sudo mdadm --create --verbose /dev/md0 --level=stripe --raid-devices=4 /dev/sd[abcd]1 [20:14] mdadm: super1.x cannot open /dev/sda1: Device or resource busy [20:14] mdadm: ddf: Cannot use /dev/sda1: Device or resource busy [20:14] mdadm: Cannot use /dev/sda1: It is busy [20:14] mdadm: cannot open /dev/sda1: Device or resource busy [20:17] Slugs_: first of all, why partitions? [20:18] Hi RoyK [20:18] well, that’s one of the guides i followed online [20:18] don [20:18] i can delete the partitions [20:19] Slugs_: and --level=stripe - is that like russian roulette [20:19] lol [20:19] yeah i know you and i discussed this, i didn’t care about data loss [20:20] partitions are good. [20:20] jrwren: nah [20:20] jrwren: using raw disks is better [20:20] why? [20:20] jrwren: helps things when you want to extend [20:20] how? [20:20] jrwren: why do you want those partitions anyway? [20:20] need everything to start offset a few MB so I can keep bootsectors at start of disk. [20:21] jrwren: hey! I have a partition! it's 2TB! [20:21] right, 2TB minus 2MB so that I can boot from any disk in the set [20:21] jrwren: don't use the OS disks for data [20:21] why? [20:21] jrwren: use data disks for data, use separate OS disks [20:21] why? [20:21] jrwren: because it makes things easier [20:21] jrwren: recovery etc [20:22] how? [20:22] lol [20:22] jrwren: but then, I've only worked with this for 20+ years :P [20:22] could it be the same year 20 times? :p [20:22] jrwren: take a good advice, please [20:22] why? [20:23] jrwren: just google it [20:23] ok. [20:23] * RoyK is not in a mood of teaching idiots how to ignore advices [20:23] ok so i have 4 disks, all 4TB a peice [20:23] used only for data [20:23] Slugs_: why raid0 [20:24] i was doing it for speed. [20:24] but i think yiou migfht have said not really good performance with 4 TB disks anyway [20:24] Slugs_: you'll get the same speed of a raid5 with 3 drives as a raid0 of 4 drives, sequencially [20:24] what do i do with the other disk? [20:24] Slugs_: and it saves a day or three if you use raid5 [20:25] when that one drive dies [20:25] and it will [20:25] yeah [20:25] ok [20:25] what do i do with the 4th disk though? [20:25] Slugs_: alternatively, you can test it both ways and measure for your specific application and use which works best for you. [20:26] yeah makes sense [20:26] Slugs_: just use a single drive, even a thumb drive, for the root [20:26] Slugs_: then use the 4TB things in raid5 [20:26] yeah i have 1 500 gb for root [20:26] and 4, 4TB drives for data [20:26] nice [20:27] Slugs_: then create a raid5 on those [20:27] Slugs_: you can change to raid0 later if you want [20:27] Slugs_: so your 1 500GB is /dev/sde ? [20:27] ok so i can do raid5 on 4 disks then [20:27] jrwren: yes exactly [20:27] Slugs_: just use raid5 on those, directly on the disks [20:27] Slugs_: no partitions needed [20:27] Slugs_: good. just making sure it isn't sda. :) [20:28] how did you know thaT? [20:28] lsscsi? [20:28] jrwren: it doesn't matter if it it's called sda or not [20:28] RoyK: it would matter if the 500GB disk is sda, and he just ran mdadm passing sda1 as a partition to use. [20:29] jrwren: it wouldn't work [20:29] Slugs_: good luck. [20:29] RoyK: exactly. and Slugs_ said, 'this command doesn't work" [20:29] specifically /dev/sda1 in use. [20:29] Slugs_: pastebin lsscsi or output of /proc/partitions, please [20:30] or both [20:30] or output of lsblk === markthomas is now known as markthomas|away [20:30] ok [20:31] i just deleted the parttions though… one sec [20:31] http://pastebin.com/bT6faShV [20:32] ok, sde is the root? [20:32] yes it has the OS on it [20:32] http://pastebin.com/4R5w304i [20:33] here is the output of lsblk [20:33] then mdadm --create --level=5 --raid-devices=4 /dev/md0 /dev/sd{a..d} [20:33] i don’t have lsscsi [20:33] no need [20:33] you can install it later [20:33] mdadm: cannot open /dev/sda: Device or resource busy [20:34] oh wait i did have this error [20:34] one sec [20:34] what does mount say? [20:34] is sda1 mounted? [20:34] not in df -h [20:34] Slugs_: i noticed the openstack vg name. are you going to run swift on that storage? [20:34] is there another way to tell? [20:35] jrwren: id like to use it as blockstorage yes [20:35] cinder [20:35] Slugs_: swift backed cinder? [20:35] Slugs_: raid0 for blockstorage? [20:35] you must be insane [20:36] i was testing out speed, it wasn’t a final configuration [20:36] just testing [20:36] not planning on using it… [20:36] Slugs_: i'll assume it is not production and not call you names :) [20:36] ;) [20:36] thanks [20:36] what does /proc/mdstat have to say? [20:36] or pvs [20:37] Personalities : [20:37] unused devices: [20:37] PV VG Fmt Attr PSize PFree [20:37] /dev/sde5 openstack-vg lvm2 a-- 465.52g 52.00m [20:37] ok [20:37] can fdisk access /dev/sda? [20:37] yes [20:37] and delete that partition? [20:38] WARNING: Re-reading the partition table failed with error 25: Inappropriate ioctl for device. [20:38] The kernel still uses the old table. The new table will be used at [20:38] the next reboot or after you run partprobe(8) or kpartx(8) [20:38] then i ran partprobe [20:38] didn’t give any errors, but i guess something is still confused [20:39] remove the partition and reboot [20:39] ok [20:39] rebooting .... [20:43] ... [20:43] it’s back up [20:44] oh boy [20:44] pastebin lsblk again [20:44] ok [20:44] or did mdadm --create work? [20:44] http://pastebin.com/aECe1zwt [20:44] no the drive names changed [20:45] good thing i didn’t just run it [20:45] well like you said it might not have letme [20:45] let me* [20:45] ok, so mdadm --create /dev/md0 --raid-devices=4 --level=5 /dev/sd{b..e} [20:46] dev names change [20:46] md uses UUIDs to determine which drives are what [20:46] it worked [20:47] then vgcreate [20:47] and lvcreate [20:47] and put a filesystem on top of the lv [20:47] the guide said , assemble [20:47] so… this is diff [20:47] * Slugs_ is looking up vgcreate and lvcreate [20:47] assemble is what you do when things mess up [20:48] i see [20:48] apt-get install lvm2 [20:48] but that's probably installed already [20:48] pvcreate /dev/md0 [20:48] it is [20:48] vgcreate mysuperdata /dev/md0 [20:49] lvcreate -l 100%FREE veryimportantdata mysuperdata [20:49] or something about that [20:50] Slugs_: what guide? [20:50] one that sucks i guess [20:50] https://help.ubuntu.com/community/Installation/SoftwareRAID [20:51] Slugs_: did you create the pv or vg? [20:51] yes RoyK just working on the lvcreate and seeing what that did [20:52] it creates a logical volume [20:52] Slugs_: server guide has this: https://help.ubuntu.com/lts/serverguide/advanced-installation.html for future reference [20:52] one you need for putting a fs on [20:52] ok [20:52] jrwren: thank you [20:52] so after creating the lv [20:52] what did you name it? what did you name the vg? [20:53] vg = data [20:53] what’s the veryimportantdata part? [20:53] just a name [20:53] descrioption/? [20:53] ok [20:54] you can rename it or remove it [20:55] hmm [20:55] root@openstack:/home/openstack# lvcreate -l 100%FREE /dev/md0 data [20:55] lvs [20:55] ? [20:55] Volume group "md0" not found [20:55] if you didn't give it a name, it'll be named automatically [20:56] erm [20:56] oh i see [20:56] Logical volume "lvol0" created [20:56] lvcreate -n myname -l 100%FREE data [20:56] lvol0 data -wi-a---- 10.92t [20:56] i can rename it then,,, ok [20:56] lvrename [20:58] what’s the path for the logical volume [20:58] root@openstack:/home/openstack# lvrename lvol0 UMS [20:58] Path required for Logical Volume "lvol0" [20:58] Slugs_: from the manual lvrename /dev/vg02/lvold vg02/lvnew [20:59] ok got it [21:00] now i guess all i need is a filesystem on it? [21:00] then put a filesystem on it [21:00] I'd suggest xfs for most use [21:00] ext4 isn't too good on large things [21:00] I spent eight hours waiting for a 8TB ext4 filesystem doing fsck two days back [21:00] mkfs.xfs /dev/mapper/data-lvol0 [21:01] oh my [21:01] mhm [21:01] ok [21:01] then add to /etc/fstab [21:01] mount -a [21:01] the normal things [21:02] ive seen so many different posts on what to put in your fstab file [21:02] I personally have misgivings about xfs but I admit it's from anecdotal experiences [21:03] Slugs_: /dev/mapper/data-lvol0 /mountpoint xfs defaults 0 0 [21:03] qman__: probably anectdotal [21:04] I think the unable to shrink it makes it not be defualt for lots of desktop but not really sure why most servers you would want to shrink a partition that much [21:04] although sparse files is a nice feature [21:04] I've lost data several times with xfs, never with ext3 or ext4 [21:04] ianorlyn: most filesystems won't be shrunk [21:04] RoyK: is mount -a supposed to just automuont this new fs? [21:04] Slugs_: yes [21:04] the only filesystem I've seen shrink was FAT, via the fips tool. scary as hell. [21:05] nothing appears to happen when i type df, or is it not suppoosed to show up there/ [21:05] Slugs_: but that data-lvol0 may be different when you changed the name of the lv [21:05] im not going to bother changing it [21:05]