=== wedgwood is now known as wedgwood_away [00:43] Seriously, any ideas how best to troubleshoot/diagnose a server being unable to resolve names? [00:44] It's perfectly fine for pure IP addresses, for the record, so I can ping 8.8.8.8 or local servers to my heart's content. [00:57] keithzg: what's in /etc/resolv.conf? [00:57] keithzg: what's in /etc/nsswitch.conf? [01:00] sarnold: the contents of both (with my company name removed, heh) are up at http://paste.ubuntu.com/5556793/ [01:02] I fear I have inherited a sprawling, interlocking and precarious beast from the previous sysadmin (the central router which ostensibly acts as one of the nameservers----it's the 10.1.186.204 one---is running Trustix, to give you an idea.) [01:03] oh my. blast from the past. [02:23] howdy room [02:23] I have a question about software raid [02:24] if you add an additional drive in a software raid 5 set does it expand the total data capability of the raid set? can you even add an additional drive? [02:25] you can and it does. See man mdadm and look for --add [02:26] thanks lifeless [02:26] you don't seem very lifeless to me [02:26] :P [02:27] mattrae, you cant easily add drives to a raid 5 type device [02:27] escott: why do you say that? [02:28] lifeless, because it is stripped (with a checksum) [02:29] mattwj2002, sorry not mattrae [02:29] escott: that just makes the software to do it complex, it doesn't make it hard to use. [02:29] no problem escott [02:29] has anyone ever done that? [02:29] escott: mdadm --add /dev/mdX /dev/sd; mdadm --grow /dev/mdX --size max; resize2fs /path/to/mountpoint [02:30] escott: the whole stack supports online resizes upwards :) [02:30] escott: see the man pages, but its most definitely easy. [02:31] lifeless, but it does require moving the data around. which is not exactly in the spirit of just adding a device and it being immediately available [02:31] escott: you don't have to wait tough [02:32] lifeless, im more skeptical of online personality switches [02:32] escott: all unmigrated data is safe to read, all new writes go into the new layout [02:32] thanks guys [02:33] may need a --raid-disks parameter in the grow, its been a while. [02:33] no problem [02:33] escott: raid5-6? Thats quite simple too; the Q parity is first calculated onto the new disk [02:33] I just wanted to know if it was possible [02:34] escott: then the whole array is walked migrating from the transient 5+Q layout to the final layout (normally left-symmetric) [02:34] mattwj2002, it is certainly possible and it can be done online, but it is a personality switch and is therefore equivalent to reading all the disk off and writing it all back [02:34] mattwj2002, im super cautious about stuff like that [02:34] I hear ya [02:35] one thing that is worth doing is adding a bitmap first [02:35] mdadm --grow --bitmap internal /dev/mdX [02:35] bitmap? [02:35] what is that? [02:35] that makes poweroff's in the middle of the rebuilds not require rescanning anything on the new disks [02:35] lifeless, i get how it works and i dont dispute there is online switching capability. its just that if i put something into raid its because i want to be really careful [02:36] mattwj2002: gives you transactional writes - writes to bitmap marking blocks as dirty, writes to the blocks, clears the bitmap. [02:36] mattwj2002: its a bit of overhead so normally off, but useful to have on when doing rebuilds or reshapes [02:37] okay cool [02:37] thanks guys :) [03:01] hey guys, if I get say, 32 GB ram, will the OS Be able to utilize all that ram for caching _everything_ after being accessed, without ever having to decache them? [03:01] as in, is there a setting somewhere to specify how much RAM is to be used for disk caching? [03:02] rurufufuss, there are tunables for the kernel caches [03:02] vm.swappiness for the win. [03:02] you can adjust it as much as you want, but default is to use all ram for cache/buffers [03:03] and even program memory :) [03:03] oh, so by default it already tries that huh [03:03] rurufufuss, well, atleast for the last 15years or so [03:03] so I guess this cheap ram I'm eyeing could be beneficial after all? reduces potential swapping and the likes [03:04] awesome [03:04] really depends [03:04] my desktop machine has 16gigs ram, I normally dont get over 8gigs, unless I start kvm up [03:04] really depends exactly what the machine does [03:05] also, is it possible to see how much space a ramdisk (mounted tmpfs) is using? [03:06] as much space as is in it, df -h [03:06] unless it has swapped it to disk [03:07] hmm, it has multiple tmpfs listing.. I thought I did unmount it though [03:07] with umount -V thedirectory/itwas/mountedto [03:07] normally there are like 3 tmpfs [03:07] there's like 13 here [03:08] hmm, no more /dev/shm [03:08] ah, it's just /run now, with symlinks [03:08] I use a lot of tmpfs on my servers [03:09] ever unmount them? [03:09] many times [03:09] so they should disappear from df-h afterwards right? [03:10] only if nothing is using it [03:10] wait, so you can unmount and it'll still "exist"? [03:11] I think I better buy that 32gb ram [03:12] these days, I see no point in not loading up ram, at $50 per 8gigs a stick [03:12] loading up new servers with 512gigs [03:13] that's true I guess, they're so cheap === aarcane_ is now known as aarcane [03:45] does anyone know if I can mount a linux RAID 1 disk with an ext4 partition as just a regular drive in a different system, and access the data? [03:45] I should say, software RAID 1. [03:46] mdadm partitions have their own metadata and uuid on the disks [03:47] if you take a single disk from a 2-disk raid1 set, you can assemble it in degraded mode and mount the fs from it, yes. [03:47] you'll want to make sure you tell mdadm to expose it readonly if you want to take it backw ithout doing a full consistency scan [03:48] OK. [03:48] if its a 3-or-4-or-more raid1 set, I don't believe you can [03:48] because they are more complex layouts [03:48] nope, it's just 2. [03:49] to elaborate a little, I have an oldish Debian system with RAID 1, and I want to switch to Ubuntu. so I was thinking I would break the mirror, install Ubuntu to one drive as a degraded RAID 1 if possible, then copy the data off the other drive before adding it into the RAID. [03:58] hello === yofel_ is now known as yofel [04:51] /win/win 13 [04:58] need build a powerpc cloud image, what's special about that image? just a network enabled rootfs will be fine? [04:59] not sure if i should use vmbuilder or lxc-creator, want to build it so i can load it under openstack/lxc === paddymahoney1 is now known as paddymahoney [10:51] Hi, just configured IPv6 on my Ubuntu server but when I try to ping external addresses it tries to ping from my Scope:Link address instead of the Scope:Global address. Does anyone know what might cause this? [10:51] And of course it then fails with: Destination unreachable: Beyond scope of source address [10:52] I can ping my IPv6 address from external sources just fine [10:56] route? [10:58] # ip -6 route 2a02:28f0:0:c::/64 dev eth0 proto kernel metric 256 fe80::/64 dev eth0 proto kernel metric 256 default via 2a02:28f0:0:c::1 dev eth0 metric 1024 [11:12] Ben64, was that what you were asking for? [11:15] dunno, i've never had a problem with ipv6 [11:15] but if you can't ping out but you can ping in, sounds like routing [11:15] had something similar with ipv4 recently === 16WAAHOCF is now known as paddymahoney [12:50] I've got a samba server running (at least it should be running) and I've installed Swat t ostart configuring it. But I get no reply in the browser when I try localhost:901 from another machine on the network (via wireless) What can I check to see hwat's wrong? [13:00] duncan-nz: 'localhost' is the machine you're on - if you replace 'localhost' with the IP (or hostname if you have DNS or a hosts file configured), that should work [13:01] IP of the *server* that is [13:01] hi thanks. first though how do I stop the ping I started on localhost ... ? :) [13:02] duncan-nz: ctrl-c? [13:03] whew! thanks. [13:06] this is where i always get confused. when i run ifconfig i get quite a few ip addresses. [13:06] i've tried the one listed as inet and Bcast but they don't work... [13:07] (gotta go out and help the kids for a few minutes) [13:11] duncan-nz: the inet one is the one that matters... is the server on the same network as the computer you're trying to access it on? [13:13] other things to consider: firewall on the client machine? firewall on the server? can the server ping the client? can the client ping the server? [13:17] well, ping localhost:901 on the server machine doesn't work, but ping localhost there works fine. [13:18] when you ask if there server can ping the client. I haven't got samba up and running with access from other machines yet, i guess that's what you mean by clients? [13:21] duncan-nz: client in this case just means "the machine with which you're trying to reach the server" [13:21] ping localhost:901 will ping the machine you're on at that time, whether that's the client machine or the server [13:22] get the IP address of each machine (inet address) and try pinging one from the other (without the ":901" [13:22] and ping localhost:901 on the server machine (SSH via Putty) gives no reaction. [13:22] ) [13:23] "ping localhost" = "ping myself" - is that what you're trying to do? [13:24] duncan-nz: this might help: http://en.wikipedia.org/wiki/Localhost [13:24] just now i'm trying to see if there is any reply from port 901 on the server. when I try to get a reaction on the server from a SSH to the server I get no reaction. I'm assuming that localhost when written in SSH is regarded as on the machine I'm accesing. [13:24] I see [13:25] so I don't think Swat is running, or maybe samba itself is not... that's what I'm tryingt ot work out. [13:26] ping doesn't work that way - you can't ping a specific port [13:26] just a host [13:26] oh, so I can't use it to test what's wrong! well that's an important thing to learn. Thanks. [13:26] you could try 'nmap localhost' (nmap may need to be installed) [13:26] ping just tests whether a host is "up" (available on the network) [13:27] nmap will show which ports are open [13:28] I have four open ports, all tcp and each for ssh, netbios-ssn, microsoft-da and port 901 for samba-swat [13:29] okay, so the port is open - now the question is can you reach the server from the other machine [13:29] I think I should do some more reading about the config files. sounds like I've overseen or misunderstood something critical. [13:30] duncan-nz: is the "client" machine a Windows machine? [13:30] I can see that the server has turned up under networks in explorer in windows on this machine. so that's something. [13:31] duncan-nz: if so, open cmd and do 'ping ip-of-server' (where 'ip-of-server' is the actual IP of the server) [13:31] client is windows. but i can jump onto a linux machine if you think that would be better for me to learn and understand this stuff. [13:31] ok [13:31] duncan-nz: I will always recommend linux over windows, but no need for what you're trying to do right this minute [13:32] that answers fine when pinged from this windows client to the servers inet ip. [13:32] okay, then try opening a browser and entering ip-of-server:901 [13:33] i'm trying to do everything over SSH and whatching what happens in this windows machine. [13:33] okay - I was just looking back at the original question you asked - you're no longer trying to access the GUI? [13:34] when i send firefox to the server I get 'can't make connection' when i try the server machine at port 901 i get 'connection reset'. so something must be reacting to my attempt at connection, or no? [13:35] I want to get into swat to configure Samba - I doubt I have configured samba correctly and thought Swat would be easier to learn and understand. [13:36] Swat is a Samba gui isn't it? [13:36] I've not used it. My samba knowledge is basically limited to getting a windows machine to see the printer hooked to my linux box ;-) [13:37] I would recommend asking in #samba, though [13:37] ok. I think i sould spend some more time here: https://help.ubuntu.com/12.04/serverguide/samba-fileserver.html working on getting the config right. [13:37] yeah - and I would read up on networking basics [13:38] yeah, i've realised I need to get some basic knowledge here. fun projekt... thanks for the help. [13:42] duncan-nz: sure ;-) [15:01] Greetings all, I'm currently setting up a new web server running Apache. I would like each file to be owned by its own owner and for scripts like WordPress to be able to edit the files. Can you give recommendations on what I need to install? [15:03] Katronix: they just need to share a common group, and that group be the web server group, that's it [15:04] ikonia so instead of it being owned by chris:chris it would be chris:nobody ? [15:08] Katronix: if nobody is the group of the webserver, sure [15:08] ikonia ok thanks [15:20] gotta love it when you ask your new hosting company for your IPs, and they say "yes you get 4 IPs" [15:22] hi guys how do i unmount an NFS share..? [15:22] like any other volumes. umount [15:23] melmoth: exact coomand.? [15:23] command* [15:23] umount /path/to/mountpoint [15:23] if i unmount it the storage file of it will be delete or not..? [15:23] no, just unmounted [15:23] ruben231 unmount just tells the server not to access it [15:24] may want to make sure your not using any files on it before unmounting it [15:28] Katronix:but i have this mount as NFS share -----------------> 172.16.0.205:/volume1/NewMP3Recordings - 5.4T 4.8T 637G 89% /var/spool/asterisk/monitorDONE/MP3 [15:28] how do i unmount that [15:28] it won't unmount if it's got open files [15:28] ruben231: umount /var/spool/asterisk/monitorDONE/MP3 [15:33] ikonia: thanks [15:54] to configure an ubuntu server for an additional IP on eth0 would you refer to it as eth0:1 ? [15:54] I've made basic changes to my smb.conf file, created a smb.conf.master file, but when I try to run $ sudo testparm -s smb.conf.master > smb.conf [15:55] it says "-bash: smb.conf: permission denied" [15:55] Katronix: just add "up ip addr add 1.2.3.4/24 dev eth0" to the end of the interfaces block [15:55] eth0:x is deprecated [15:56] ... that should have ended with "... >smb.conf" [16:04] Roy OK, what is the newest way to start networking? [16:06] or restart that is [16:07] sudo restart smbd [16:08] (if it's samba you're talking about) [16:08] duncan-nz no I was refering to networking [16:08] bummer, for 2 secs I thought I'd understood something... [16:09] but I have just learnt that restarting services is no longer done with init scripts. [16:09] yeah I just saw that too [16:09] I've tried service networking restart and it says it doesn't know what is meant by stop and then says its stopped lol [16:11] what do you mean by networking, surely there's a specific package to call? [16:11] or is 'networking' a package? [16:11] added a new IP to the server, networking at least used to be the way to tell it to re-read its configuration for what IPs it had [16:12] so we're two people looking for help talking to each other? bummer again. [16:12] lol yep guess so seems everyone else are off doing other things [16:13] so you know Samba [16:13] ? [16:13] a little, been a while since I used it [16:15] I get why I need to restart smbd, but what is nmbd - how do I get info on what something is in bash? [16:15] nmbd is something that works with smb [16:16] hi guys im having hardtime...finding what cause the huge storage of my server HDD, any idea how to find it..? -----------> http://pastebin.com/1W9ct6du [16:17] ruben231 you don't know what is causing: /home/openweb/.Private ? [16:20] Are you still here RoyK? [16:21] yep [16:22] RoyK so this should work: http://pastebin.com/VBVMtPiD ? [16:25] I would have written it as this http://pastebin.com/DKtzwTvm [16:25] but then - keep in mind that those two addresses are on different subnets [16:27] Katronix: $HOME/.Private is a symlink used if you encrypt the homedir [16:27] OK and then after that what command do I run to get ubuntu to reconfigure the network? is it service network restart? [16:27] networking [16:28] ok === paddymahoney1 is now known as paddymahoney [16:29] service networking restart does not seem to have the desired effect :( [16:29] /etc/init.d/networking restart [16:29] perhaps [16:30] btw, can you manually do 'ip addr add 192.95.23.246/24 dev eth0' ? [16:31] you mean type that in at a bash prompt? [16:31] yes [16:31] 'up (something)' in interfaces means 'when this nic goes up, run this' [16:31] RTNETLINK answers: File exists <<< was the response [16:34] so what does ip addr list tell? [16:36] http://pastebin.com/DUpgX4pP so I guess it is there? even if ifconfig doesn't list it? [16:36] no, ifconfig won't list more than one IP per nic [16:37] unless you use the older ethx:X [16:38] ok, so going by that it should work if i ping it? [16:39] it should [16:39] yay it works [16:39] but clients on 192.95.29.0/24 probably won't be able to reach it unless you setup routing correctly [16:40] I didn't have any control of the IP, it was the host who assigned it [16:47] so the one on 192.95.29.x is from dhcp? [16:48] they are both from the company I'm getting the dedi from [16:50] thanks for the help all, time to go off and do other things === paddymahoney1 is now known as paddymahoney === paddymahoney is now known as Guest95608 === Guest95608 is now known as paddymahoney [17:02] for your weekend amusement: a venn diagram of unix commands/utils, us states and territories and irish license plate codes: http://goo.gl/Zy3eY [20:24] I do " sudo service dovecot start" service seems to start, i have enabled logging in docecot.conf, service seems to instant die, and my logs are empty. any advice? i have tried to touch the logfiles and gave them chmod 777 to amke sure dovecot could write them if the user\group or anything like that should be wrong. [20:25] I am logging to info_log_path = /var/log/dovecot-info.log log_path = /var/log/dovecot.log === gary_poster is now known as gary_poster|away [21:27] Hey Guys. What can i do if i accidently blocked port 22? [21:29] login on the console and unblock it [21:29] ssh is on port 22... [21:29] that's why I suggested the console [21:29] what console do you mean? [21:30] The physical keyboard [21:30] its a root... [21:30] i dont have physical access [21:30] or a serial remote console or remote management console if you have a real server in there [21:30] the only thing working is port 80 and the CP [21:31] Styler2go: So no telnet either? [21:32] only port 80 and the cp seems to work [21:33] what options do i have? [21:34] Probably depends on your control panel options there. [21:34] a restart wont help anything, or? [21:34] its a parallels plesk power panel [21:35] Styler2go: Did you turn off/remove sshd, or did you just do something like make an iptables rule which locked you out? [21:35] ufw [21:35] but i made ufw 22 allow... [21:37] i could uninstall ufw.. would thi shelp? [21:40] Styler2go: I'd probably try that, yes [21:41] ok it got removed successfully... [21:41] semms not... [21:46] Styler2go: Since there is probably a way to re-enable ssh access from Plesk, I'd suggest to call your site provider tech support. I do see a page explaining how to let ftp users access by ssh here http://kb.mediatemple.net/questions/914/Enable+SSH+Terminal+for+FTP+user+within+Plesk#dv_40 [21:47] But they would know more about the inner workings of that control panel [21:47] grr [21:47] ok i have to think about it [21:49] ( and, it's not really a support question for ubuntu server in the sense that it's something you did within the control panel which is the issue, and not the underlying operating system is having some issue) [21:49] i also cant restart the server.. [21:50] the guys in #ubuntu told me to go here [21:51] ok i got the restart working, i have access to ssh back... [21:52] Styler2go: Good, the reboot would have flushed the iptables rules [21:53] maybe you could help me now with the real ubuntu... i want to set up a firewall to block if there are coming more than X/sex requests from one ip [21:55] Styler2go: I have to go to a hockey game shortly. But someone else might take it up. If you just want to slow down people trying to brute-force password attack maybe look at fail2ban though [21:55] https://help.ubuntu.com/community/Fail2ban [22:22] <_KaszpiR_> yawn [22:23] <_KaszpiR_> I'd like to install ubuntu LTS minimal ion UEFI via usb [22:23] <_KaszpiR_> any tutorial? [22:43] <_KaszpiR_> ugh now I know, the md5sums got borked [22:54] <_KaszpiR_> ok, looks like it is working [23:06] <_KaszpiR_> whehe, works [23:15] I have a RAID5 array which used to consist of three 2TB drives. Last week I added another drive and after that I grew the array to include four devices. Although, I forgot to modify the filesystem size. I will attempt to do this overnight, but I'm wondering if this command is correct: "mdadm --grow /dev/md0 --size=max"