=== edwardly_ is now known as edwardly [00:06] I'm trying to add NIS extensions to my samba4 domain, following the instructions here: https://wiki.samba.org/index.php/Using_RFC2307_on_a_Samba_DC -- Checking my environment, I don't have them installed, and it says to use ypServ30.ldif to add them. However, this file does not exist anywhere on my system. Running Ubuntu 12.04 with Samba 4.0.0~alpha18.dfsg1-4ubuntu2 === Lcawte is now known as Lcawte|Away [00:30] here's my /etc/exports, but when I try to connect to my shares from OSX 10.9, it's slow to connect. Any idea why? http://pastebin.com/raw.php?i=9hPvZDgX [00:53] I have "rw" in my nfs share, but I can't write to the share - how do I enable the ability to write to it? [01:01] fridaynext: file permissions, most likely - that's a really bad configuration from a security standpoint [01:02] qman__: do you have any links to tutorials about how to increase the security without spending four hours reading the man pages? [01:02] it's very simple [01:03] 192.168.1.0/255.255.255.0(ro,async,no_subtree_check) [01:03] is what I have on mine [01:03] huh.. [01:03] qman__: and how is mine less secure than yours? [01:03] to make it work, your UIDs have to match on all affected systems [01:03] and the users have to have permissions to the files [01:04] that's why i set the uid to 1000, since I can't change my UID on OSX. [01:04] fridaynext: yours turns off all the security features that are built in, allows root and guests to write to anything, etc [01:04] I want to be able to write to all the folders I connect to - but I can't seem to make that happen without adding all the extra variables. [01:05] change your UID on the linux side then, or add a specific user for it [01:05] then set the ownership and permission on the files [01:05] I tried doing that with groups before, but I guess GID is not enough... [01:06] NFS is designed to be used within a unified environment, where everyone authenticates to a central authority, UIDs are synced, etc [01:06] without one, security is impossible [01:07] but allowing root and guests is a particularly bad idea [01:09] gotcha - i'm making those changes now to see if it works. [01:10] qman__: so since you have 'ro' in yours - is that read only for even your UID? [01:10] yes, for that one, I have an rw one too [01:10] it's the same except for that [01:11] I also have a samba share on the same location as the rw nfs share, all the files have enforced group-ownership through several methods, and all the share users are a member of that group [01:12] see with your settings, I can't connect to the NFS share. Even with creating a user on my *nix box with the same UID as my OSX user. [01:12] the user has to have at minimum read permission on the directory you're sharing, write to write to it [01:13] and execute on directories to list them [01:13] the user does. [01:14] the 'casey' folder is the one i'm trying to share: http://pastebin.com/raw.php?i=SJZwXUZa [01:14] and my nfs export line is: /media/raid/tm_backup/casey 192.168.0.0/24(rw,async,no_subtree_check) [01:15] ok [01:15] I gave 'node304' the same UID as my OSX user I'm logged in as. [01:16] trying to find where the log is stored [01:17] I can't find it either... [01:18] also, are you running nfs-kernel-server? [01:18] qman__: ye [01:18] s [01:18] ok [01:20] try changing to /255.255.255.0 instead of /24, just to test [01:20] I'm not certain if it supports CIDR or not [01:21] i can connect to another of my shares with /24, but I'll test this anyway. [01:21] ok [01:22] what error, if any, do you get trying to mount it? [01:23] https://lh5.googleusercontent.com/--VrYk3lnOYg/U_qP9qN8hCI/AAAAAAAAHDk/UbT4XBCdjok/s800/Screen%2520Shot%25202014-08-24%2520at%25208.22.44%2520PM.png [01:25] that error message doesn't help much [01:25] try this on the server: sudo mkdir -p /tmp/nfsmount; sudo mount -t nfs 192.168.0.X:/media/raid/tm_backup/casey /tmp/nfsmount [01:25] replacing X with your server's IP [01:26] if it mounts that at least means the NFS configuration works, and points us to a network issue or a mac specific issue [01:27] just issued that command - it's hanging. [01:28] nope - did it. looks like it's mounted. [01:28] 192.168.0.100:/media/raid/tm_backup/casey 5858043904 1247198208 4610845696 22% /tmp/nfsmount [01:29] ok [01:30] try the same thing on the mac [01:30] from the terminal [01:31] "Operation not permitted" [01:31] http://pastebin.com/raw.php?i=yfZh4n9k [01:34] try this: sudo mount -t nfs -o resvport,rw 192.168.0.X:/media/raid/tm_backup/casey /tmp/nfsmount [01:35] okay, that worked nearly instantly. [01:35] so now I need to know how to do that via the Finder. [01:35] where'd you find that fix? [01:35] ok, that tells us exactly what the issue is [01:35] nice! what is it? [01:35] that option tells the mac to use a reserved port (sub 1024) [01:35] so we either need to tell finder to do that, or tell linux it's okay if you don't [01:36] I wonder if I can add the port to my finder connect comand [01:36] actually, I don't think I can do that in my finder command. [01:36] so what's the option in /etc/exports to remove the port specificity? [01:37] insecure [01:37] add that to the list and it should work [01:38] Ah, I did that once before, and it caused the Finder to freeze up. [01:38] I just need some program that runs scripts on startup for OSX, then I can avoid this Finder nonsense. [01:38] you can configure it to auto mount I think [01:38] Yeah, finder has frozen again (once I've added 'insecure' to the list) [01:39] http://apple.stackexchange.com/questions/102092/mount-linux-nfs-share-on-mountain-lion-10-8-5 [01:39] ah - it worked with the finder after about 40 seconds [01:39] super slow. [01:39] finder might be trying to do something silly [01:39] i'm sure it is. [01:40] you've been a great help, qman__! [01:41] although, I do'nt have that same file that the poster is talking about on that page. [01:41] no /etc/auto_nfs in my 10.9 build. [01:46] this may or may not help: http://www.techrepublic.com/blog/apple-in-the-enterprise/mounting-nfs-volumes-in-os-x/ [01:46] unfortunately this is getting pretty deep into mac-specific territory I'm not familiar with [01:47] yeah, that article was written bakc when you could mount NFS with disk utility. Apple removed that option now. [01:47] heh [01:47] yup [01:48] well, I think there's a feature of launchd to make stuff run at startup that could be scripted, but that's getting pretty deep for a filesystem mount [01:48] oh - that did seem to work [01:49] I just wasn't looking in the right place. [01:49] for the performance problem, it could either be using poorly optimized defaults for the block sizes and such, or it could be trying to do something silly having to do with the apple-specific dotfiles [01:49] oh i have great performance on NFS [01:49] 100-110MB/s [01:49] pretty much maxing out my RAID5 array [01:50] Does that guy add too many attributes to the connect? the stackexchange post? [01:50] or do those not affect the security? [01:50] you don't need resvport since you changed the server side to be ok with it [01:50] the rest are performance-related [01:51] you can view what each one does in the nfs man page [01:51] he is missing the 8 on the rsize [01:51] they're supposed to be the same number in that example [01:54] while I'm not sure how it behaves on a mac, using "soft" on a linux client is important especially if your network isn't perfectly reliable; the default "hard" will cause any process with a file open on the share, should the share drop off the network, to freeze until it comes back [01:55] with soft, it just generates an error and allows execution to continue [02:03] good god, now I can't get it to work. [02:11] I've NFS exported my root filesystem (from host/server blue to client orange), which of course includes /mnt/ which itself mounts a secondary drive; i.e. it's a different filesystem than / So orange sees the mount as /mnt/blue/. Although I'm able to see the directory tree under /mnt/blue/mnt/ which has additional subdirs (i.e. /mnt/blue/mnt/a/b/c/d/e/f/g/h/) the files that should be under ../h/ are not there. server is ubuntu 12.04, client is [02:11] ubuntu 14.04 [02:14] any advice on whether to use ufw or just use iptables directly? [02:15] sarkis: either is fine, it's up to your preference [02:15] sarkis: ufw does a lot of things for you but can't handle every configuration possibility [02:16] ya i lean with using iptables so if i ever switch distros i don't rely on a ubuntu only thing [02:17] personally I prefer to write a shell script that does all the iptables stuff and put it in if-up.d [02:17] for anything complicated, at least [02:34] qman__: well, I can at least mount my NFS share - even though I can't get it to automount yet - but that's all I need for now, so I Can start taking backups of this machine. [02:34] Thanks for all your help tonight! [03:24] hey all, is there a way to get a list of which security updates are available for a server with aptitude or similar? [03:52] superspring: i guess one cheesy way would be to enable only the security pocket and see what comes down [03:52] pmatulis: i see, no one-liner commands to do it? :p [03:53] superspring: there probably are [03:53] superspring: hint, all security updates will have a '.#' tacked on to its version number [03:54] superspring: the kernel is at least one exception to that rule [03:55] as each kernel is its own package, not an upgrade [07:21] Hi all, ubuntu server running on the local network keeps losing ethernet connection, I've put the output of ping on pastebin. http://pastebin.com/WGe0Eq0Q How can I fix this? [07:23] trijntje: Run a packet capture and see what happens? [07:25] lordievader: how would I do that? I dont have a lot of experience debugging network problems [07:27] trijntje: "sudo tcpdump -i eth0" Hopefully it doesn't grow too large before it happens. Wireshark doesn't like large files. [07:29] lordievader: it only happens when I'm using torrents, so it might grow pretty fast [07:30] trijntje: Ah, then the problem is likely that you open too many TCP connection and DOS your router. [07:30] trijntje: Tell your client not to open a million tcp connections ;) [07:31] lordievader: I thought that as well, but the router is fine and the other pc's attached to the router are also fine [07:32] Still, limit the number of connections and see if the problem persists. [07:35] lordievader: I thought I did, this is the relevant part of my rtorrent config http://pastebin.com/br8Fa08d [07:35] also, netstat shows that rtorrent only has 16 open connections [07:37] trijntje: Hmm, that should be fine indeed. Carry on with the packet capture ;) [07:41] ok, I added a torrent and its dropped of the network now [07:42] I can still connect to the router and other pc's on the network [07:50] lordievader: ok, its back online now and I have the tcpdump file, what should I do next? [08:02] trijntje: Analyse it with Wireshark. See if there is something fishy going on. [08:06] lordievader: "tcpdump.txt isnt a capture file in a format that wireshark understands" [08:07] hm, looks like I have to rerun it with the -w flag to make it readable for wireshark, i'll have to capture it again [08:08] thanks for your help so far, i'll be back when I got the data [08:10] trijntje: Oh, right. I allways forget that -w flag.. Sorry 'bout that. [08:16] lordievader: no problem, its not an acute problem, just anoying [08:31] trijntje: I get that ;) [08:53] are there any microsoft office irc channels? [09:13] Vladimir_: If there are, we'd be some of the least likely to know about them. [09:18] soren: thanks, no problem=) hehe === Lcawte|Away is now known as Lcawte [09:40] so, lets say I have two servers running Ubuntu Server, can I simply copy all relevant postfix files to the new machine ? [10:21] SrRaven-work: depends on what your intent is. in anycase, make sure that they don't ultimately have the same answer for $myhostname [10:22] though it defaults to /etc/mailname, so you should be ok [11:42] is enabling automatic update installs safe/best security practice [11:42] am using 14.04.1 lts [11:44] nevermind bad packages can wreck havoc, so i'll do manually [13:07] Is there anyone who use Rancid ? === rapa is now known as Reidadolbuso [13:07] Maybe can help me set it up, I'm having alot of errors [13:24] I get one problem with Rancid, when I run clogin to a switch it says that the .cloginrc file does not exist in the /home/rancid/.cloginrc directory. Why does it look in that directory since I created that file in /var/lib/rancid/.cloginrc ?? very strange, how do I change where clogin looks for that file? [13:26] <_ruben> wonder if i can somehow combine bonding and mpio on the same set of nics (as in: bonding to have 1 management ip reachable over all nics, mpio to provide extra bandwidth for iscsi) [13:27] Vladimir_: is there a GUI for rancid? If so, look to settings [13:27] pgar23: no I have to get it to work before I can get web gui to work :/ [13:42] <_ruben> seems setting fail_over_mac to active might work for me === Vigour_ is now known as Vigour === Tzunamii_ is now known as Tzunamii === balloons_ is now known as balloons === hachre_ is now known as hachre === DavidDuffey is now known as dduffey === psivaa_ is now known as psivaa === hxm is now known as Guest94481 === tjaalton_ is now known as tjaalton === croepha_ is now known as croepha [15:14] hallyn, when we move to systemd does that magically solve the issue with upstart and network namespaces ? [15:15] without adding other issues? [15:18] smoser: i think so. don't know enough about systemd [15:19] systemd not add other issues? :) [15:23] patdk-wk, ewll, those will be magiclaly solved === Beliq____ is now known as Beliq [15:23] with pixie dust [15:24] ah, my daughter is good with that, I'll ship her over === JanC_ is now known as JanC === matsubara is now known as matsubara-lunch === pleia2_ is now known as pleia2 === a1berto_ is now known as a1berto [16:52] apb1963: what was the missing piece? [16:52] Is it possible to create a deployment server on a minibook? [16:52] sarnold: firmware [16:52] apb1963: groan. [16:53] apb1963: nice debugging :) [16:53] well, I can't really take the credit.... TP-Link support suggested it. [16:54] (after suggesting some other somewhat dumb things that might have made sense for some people but was clearly a checklist of items that the support girl had no idea what she was really suggesting) [16:55] I have two new issues though :/ === Ursinha is now known as Ursinha-afk [16:55] It just never ends.... [17:11] apb1963: oh, nice. I guess it makes sense for them to include firmware update in their list of suggestions [17:11] well, that was last on the list [17:12] sarnold: but hey, I'm just glad it's working... better ;) [17:12] :) [17:13] sarnold: how's your NFS? [17:13] apb1963: heh, I haven't used nfs in a decade. [17:13] sarnold: yeah, same here... and now I need to [17:14] sarnold: unless of course you can think of a better solution [17:15] apb1963: well, depends what you're trying to do. the new hotnesses are iscsi or ceph, but I like the simplicity of NFS. [17:15] sarnold: never heard of those (this is why I ask these things). I just want to share my filesystem with my other computer, same LAN. [17:16] apb1963: NFS is probably best then [17:17] sarnold: if only it was working :/ [17:17] sarnold: I've got the strangest problem. [17:17] not surprised :D [17:17] apb1963: firewalls on either system? [17:18] sarnold: yes... but I don't think that's the issue [17:18] sarnold: 103 7988 ACCEPT all -- any any 192.168.0.0/24 anywhere [17:19] I've NFS exported my root filesystem (from host/server blue to client orange), which of course includes /mnt/ which itself mounts a secondary drive; i.e. it's a different filesystem than / So orange sees the mount as /mnt/blue/. Although I'm able to see the directory tree under /mnt/blue/mnt/ which has additional subdirs (i.e. /mnt/blue/mnt/a/b/c/d/e/f/g/h/) the files that should be under ../h/ are not there. It DOES find /mnt/blue/etc/foo; just [17:19] not /mnt/blue/mnt/files - server is ubuntu 12.04, client is ubuntu 14.04 [17:20] As if that wasn't enough, at the moment the client is giving me: clnt_create: RPC: Program not registered when I do a showmount -e [17:23] apb1963: nfs export will only export that one filesystem; if you want clients to have submounts in the proper location, you've got to export that filesystem too, and mount that filesystem too [17:24] apb1963: program not registered is probably missing the rpc stats or whatever arranges the dynamic ports that nfsv1 loved so much [17:24] portmapper? [17:24] Actually I tried mounting just /mnt [17:24] well, programs that use portmapper [17:24] like lockd, statd, mountd, ... [17:25] ah, thanks [17:25] patdk-wk: that still begs the question.... how do I fix it? [17:26] using nfsv4 actually [17:26] nfs4 has lots more things [17:26] like id mapping and stuff [17:26] nfs4 is a whole nother level of difficulty, over nfs3 [17:26] actually... I don't care what version I use, as long as it works [17:26] and 10.04 doesn't support nfs4, I don't believe [17:26] I would be supprised [17:26] 12.04 here [17:27] I thought that was 12.04 [17:27] oh, client is 14.04 [17:27] thought that said 10.04 [17:27] server 12.04, client 14.04 [17:27] did you edit /etc/defaults/nfs* [17:27] to enable nfs4 stuff? [17:27] I did not [17:27] configured idmapper [17:28] and all that other nfs4 config info [17:28] nfs-common and nfs-kernel-server [17:28] and /mnt/blue isn't valid for nfs4 [17:28] I took a look at them last night... I didn't see anything that needed changing [17:28] all nfs4 exports must happen from the single nfs4 export mount [17:29] well, last I checked [17:29] idmapper and gss aren't enabled by default [17:29] and idmapper is required for nfs4 [17:29] gss is for kerberos only I think [17:29] though if your doing nfs4, you should be using kerberos [17:29] yes, gss was causing a problem... I blacklisted the module [17:29] otherwise, there is not much point of using nfs4 :) [17:29] ah [17:30] well, it's just on my local lan [17:30] so let me go back to nfs3 since I don't need the new features [17:30] oh I thought nfsv4 brought tcp support which was nice because tcp [17:30] no [17:30] that was nfs3 [17:31] nfs4 brings user authenication and mapping [17:31] though, user authenication requires kerberos [17:31] and usermapping is still broken [17:31] so yournames need to match and user id's [17:31] whoops... I already did go back to nfs3 earlier :) [17:31] sarnold I switched to tcp for netapp [17:32] now I'm not on netapp I switched back to udp, and got more benifits [17:33] patdk-wk: oh, what do you prefer about the udp service? [17:33] it doesn't have all the performance issues tcp has :) [17:33] my etc exports: / 192.168.0.101(rw,sync,no_root_squash,no_subtree_check,fsid=0) [17:33] tcp is nice, if your not staying on a local network, or if you have dropped packets [17:34] I guess I should lose the fsid=0 if I'm going back to v3 [17:36] patdk-wk: so there's really nothing to do on the client other than mount the fs, is there? [17:37] and install the nfs-common package [17:37] yep [17:37] specify vers=3 as a mount option [17:39] 192.168.0.12:/ /mnt/blue nfs auto,noatime,nolock,bg,nfsvers=3,intr,tcp,actimeo=1800 0 0 [17:40] root 407 2 0 00:37 ? 00:00:00 [rpciod] [17:40] root 415 1 0 00:37 ? 00:00:00 rpc.idmapd [17:40] root 635 1 0 00:37 ? 00:00:00 rpcbind [17:40] statd 709 1 0 00:37 ? 00:00:00 rpc.statd -L [17:40] seriously? your exporting root? [17:40] it's local LAN [17:40] still, it's odd :) [17:40] it's just (in theory) easiest [17:41] I didn't want to mess around with exporting one by one what I might need later. [17:41] one export... done. [17:41] again... in theory. [17:41] so do those look like the proper set of processes? [17:42] for what? [17:42] client or server? [17:42] client [17:43] same as mine [17:45] and the server... [17:45] root 14707 1 0 01:38 ? 00:00:00 /usr/sbin/rpc.mountd --manage-gids [17:45] root 14826 1 0 01:38 ? 00:00:00 rpcbind -w [17:45] statd 14840 1 0 01:38 ? 00:00:00 rpc.statd -L [17:45] root 16293 2 0 Aug24 ? 00:00:00 [rpciod] [17:45] root 16303 1 0 Aug24 ? 00:00:00 rpc.idmapd [17:46] well, I don't have a 12.04 nfs server anymore [17:46] but it seems right [17:46] no firewalls? [17:46] on the server yes [17:46] and everything is allowed for nfs? [17:46] or for the client? [17:46] you know what ports those rpc processes are using? [17:46] 53 4250 ACCEPT all -- any any 192.168.0.0/24 anywhere [17:47] that line doesn't mean anything [17:47] it doesn't? [17:47] what table is it in? [17:47] wihtout context, no [17:47] iptable [17:47] iptable is a program [17:47] INPUT table === unreal_ is now known as unreal [17:59] well that's interesting.... showmount -e with a hostname works... but without gives me clnt_create: RPC: Program not registered [18:00] so I guess... it's not an issue to worry about maybe [18:00] that's from the client === Guest94481 is now known as hxm [18:01] the server just hangs with the same command. I guess I'll chalk it up to different versions. None of that is my real problem... just thought maybe it might provide a clue. [18:01] * apb1963 puts away his red herring [18:02] you did do a, showmount -e 192.168.0.12 [18:03] showmount -e works for me, on 14.04 and 12.04 [18:03] patdk-wk: it's all very strange. I can see /mnt/blue/mnt/swapfile but /mnt/blue/mnt/backintime/whatever doesn't exist [18:03] well I did once I realized it needed a hostname [18:04] did you cross filesystems? [18:04] nfs doesn't support crossing filesystems [18:05] the server has /mnt/backintime mounted [18:05] its a different disk [18:05] but you exported / [18:05] not /mnt/backintime [18:05] right [18:05] nfs depends on inodes to keep things straight [18:06] when you cross filesystems, inodes get all screwy [18:06] and things don't work [18:06] this is why nfs4 has the fsid option [18:06] and cause it only has 1 export [18:06] hey folks, i just upgraded to ubuntu 14 lts from 12 lts, i'm trying to start up one of my apps and its giving me the following. http://puu.sh/b7xda/2676f0628e.png === matsubara-lunch is now known as matsubara [18:06] i suspect i need to downgrade perl from 5.18.2 to 5.14.0 [18:07] xibalba, oviously you installed some non-ubuntu software on there [18:07] that is made for perl 5.14 [18:07] correct, zimbra collaboration suite [18:07] so when you upgraded, and got perl 5.18, it's broken [18:07] can't help you there [18:07] why not just upgrade zimbra? [18:07] yup, can i downgrade perl? [18:08] i'm on the latest version...i think, double checking [18:08] yeah 8.0.7 is the version i'm on [18:08] and is the latest [18:08] never upgrade, till your software your using, supports it :) [18:08] well yeah i just made a stupid mistake [18:09] i wanted to try the do-release-upgrade, i hadn't tried it before and just jumped the gun [18:09] can i downgrade perl? [18:09] it's not supported [18:09] dunno what would happen [18:09] you could attempt it [18:10] i'd just have to compile and install [18:10] but most likely, the result is, your system will be unusable [18:10] fudge, i wish i snap shotted this thing === Ursinha-afk is now known as Ursinha [18:11] * xibalba kicks /me [18:11] The takeaway: never upgrade without a rollback plan [18:11] yeah, i wouldn't in production [18:12] this is just my personal mail server [18:12] You could upgrade to the Zimbra beta, it's available for 14.04 [18:13] dasjoe, ty didn't see that [18:20] patdk-wk: sorry... phone rang [18:20] patdk-wk: well, I thought I had tried what you said.... but apparently I didn't do it right because I just tried it now and it's working. [18:20] patdk-wk: Thank you!!! [18:42] dasjoe, thank you! beta did the trick [18:44] got lucky there, hopefully there's no show stopping bugs === Ursinha is now known as Ursinha-afk === Ursinha-afk is now known as Ursinha [20:45] Hey guys. I have kind of a wierd issue. I created a AWS 14.04 instance. When I make a sudo call I get the message "sudo: unable to resolve host (IP Address)". Its only for sudo calls, and I haven't edited the hostfile. Anyone have any ideas? [20:45] hostname file* [20:47] zzxc: Is the hostname in /etc/hosts correct (the localhost entry I mean) [20:47] ? [20:50] Yep. 127.0.0.1 localhost then the subsqent ipv6 stuff as well [20:50] lordievader: Sorry for the delay. I had to kick off a build on another machine, [20:51] zzxc: Ok, what hostname is in /etc/hostname? [20:52] ip.170.30.xx.xx [20:52] zzxc: Hmm, strange hostname. Anyhow, add it after localhost in /etc/hosts (unless it is defined elsewhere in that file). [20:57] lordievader: Its pretty common fomrating for aws instanse. [20:58] instance* One sec. I'm going to test if that worked. [20:58] Yeah. That worked. [20:59] zzxc: ;) [20:59] Funky. I've never had an instance where it didn't add that in the first place. Cool thank you lordievader. [20:59] zzxc: No problem ;) [21:00] Also the whole xx.xx at the end was just because I didn't want to post the machine's ip (though it is interal) on IRC. [21:01] zzxc: I know ;) [21:02] Figured you would. =]. Also on a unrelated note, I love vundle occationally. === phaidros_ is now known as phaidros === mwhudson_ is now known as mwhudson === TDog_ is now known as TDog === NomadJim_ is now known as NomadJim === Lcawte is now known as Lcawte|Away [23:59] so after I open up a few ports for some of my programs to use ... do I need to restart something to facilitate that? $ sudo ufw allow 4506 ... then I check it with $nmap -v 127.0.0.1 -p 4505 ... and it says its closed