400 Bad Request

[00:04] stgraber: ido, but can't right now. will try it later tonight or tomorrow [00:05] hallyn: ok, thanks [00:05] (if you want to push now, i won't mind...) thanks, ttyl [00:09] Hello all [00:10] I'm running a bunch of per-user PHP5-FPM pools and for some reason they're dying at irregular intervals. [00:10] I don't see anything around that time in syslog [00:10] Any ideas? :\ [00:10] josephoenix: what release of Ubuntu / PHP5 ? [00:13] php5-fpm is 5.3.6-13ubuntu3.5 [00:13] Ubuntu 11.10 [00:16] Hmmm it was at 6:30 am and there's a cron job running around then.. gonna look into that. [00:18] Nope, the user crons are only @reboot and @monthly [00:19] josephoenix: perhaps there is a bug in php5-fpm .. I'd start logging stderr/stdout. [00:19] josephoenix: also if you haven't disabled core files, try to see if there are any core files from segfaults [00:19] It's about 20 different instances of php5-fpm with a worker or two each that all disappeared around the same time [00:21] If core files are enabled by default, I should be able to get to them. Where do they live, again? [00:23] (Nothing in the php5-fpm log, though there might have been something written to stderr I don't know about) [00:34] I got a disk that fdisk -l says "system: linux" - how can i make it NTFS? mkfs.ntfs -q /dev/sdb1 dosnt seem to work - it just hangs [00:59] ohh cool [00:59] there is a server support [01:01] ls -al [01:01] ah sry [01:01] hello all i need help with updateing drivers for my Ethernet card but do not know the command to do this why this is, its cuz i have installed the server on my laptop then transferd it to my desctop since my desktop was unable to load the instalation [01:11] thats how you update l?s-al [01:13] is here ppl or is this chat dead??? [01:15] Pushing to ~ubuntu-server-dev/nova/essex/ without review, as it's blocking CI [01:21] New bug: #929127 in nova "linux_net.ensure_metadata_ip() fails on precise" [High,Fix committed] https://launchpad.net/bugs/929127 [01:23] hello all i need help with updateing drivers for my Ethernet card but do not know the command to do this why this is, its cuz i have installed the server on my laptop then transferd it to my desctop since my desktop was unable to load the instalation [01:32] Looking for help on unexpected soft RAID behaviour after replacing failed disk in RAID1 device. [01:34] 1. Disk failed. 2. Power down. 3. Replace disk. 4. Power up with degraded array. 5. fdisk. 6. mdadm -add. 7. Array syncs. 8 /proc/mdstat looks good. 9. Reboot. 10. Back to degraded array!! [01:34] ...any hint? :) === mathiaz_ is now known as mathiaz [01:49] Hey guys, this isn't ubuntu specific but this seemed like a good place to ask. I'm a systems engineer in a growing firm and I wanted to know, whats the best way to handle passwords more then one person should have where you can't have seperate passwords like, for example, most our machines use openLDAP for single sign on but then there is a the root password which, as we're expanding I feel I now need to share [01:50] like I imagine there is some software someone wants to recommend as a central web based password store or well, I'm jumping to conclusions but I'm sure there is some best practices industry accepted method of doing this [01:51] btw, I mention root in the event a machine doesn't connect to the net and someone needs to go to the data center and manually configure it === jetole_ is now known as jetole [01:53] jetole, well, given physical access you typically have root anyway [01:54] cwillu_at_work: so bob the admin I work with needs to go the data center and log in to server42 to find out why it's not connecting to the net and needs the root password. How does he get that root password and where is it stored? [01:55] he has physical access? [01:55] sure [01:56] bob and server42 are hypothetical but I'm trying to find a methodology to apply to our real world situation [01:56] (sorry, sec) [01:56] maybe he needs access to switch12 which doesn't support radius or multiple users [01:56] jetole, how do you secure physical access? [01:57] (i.e., why can't they just reboot, and bring up the recovery console?) [01:57] cwillu_at_work: well the data center which is one of the largest in the world takes care of that. Our cabinet is locked and the keys never leave the site. We provide ACL's to the data center of whom has access to what [01:58] jetole, then why not just let that be your local auth? [01:58] i.e., console user has root privileges [01:58] cwillu_at_work: well in most cases they can though I think ubuntu 10.04 which is our primary OS has password issues with boot into single user more (I think, not positive) but then what about switch12 for example [01:58] single user mode will permit a root login without password, it's just a config option [01:59] cwillu_at_work: I'm sure there has to be lots of other circumstances I can't think of at the moment to make this a legitimate question. no? [01:59] switch12 is another matter, but I fail to see how this really relates to that :p [01:59] cwillu_at_work: well the question is how do I store passwords where it's not an individual users password [02:02] I guess I just feel, and maybe I'm mistaken but, I feel, some things need to be shared but I need to implement some form of ACL as to who has access to those things [02:02] one approach would be to have a audited and secure database of server root passwords [02:02] index cards in envelops in a vault, for instance :p [02:02] cwillu_at_work: I considered that and was part of the conclusion I jumped to but if so, I hope someone has already designed a package to manage this [02:02] ha [02:03] you could also do that plus an otp system [02:03] a little too 18th century for my taste at index cards in vault [02:03] otp? [02:03] one-time-password [02:03] either a pregenerated list which have to be used in order (thereby making it obvious from an audit standpoint), or any number of more sophisticated approaches [02:04] the list approach has the benefit of being quite practical with index cards [02:04] hmmm. Don't know how I would do that. I know ssh has a mechanism for that but don't know how that would be applied to a physical system or then again other objects like I know we have a few switches this would apply to [02:04] (and don't scoff at that, index cards still work when the password server blows up [02:04] hmmm. good point at list and auditing [02:04] well another big project is we're eliminating all single points of failure [02:05] HA throughout the house [02:05] but I like it as far as auditing is concerned [02:05] just remember that physical access == root access [02:05] also, can you think of another good room for me to ask about this in? [02:06] the most you can do is encrypt the data such that one _has_ to be there with a password to unlock it [02:06] cwillu_at_work: yeah I know. It's not all nessecarily physical access. Last night I had to give another engineer the windows admin password so he could log into a system that wasn't playing well with active directory and fix that and the system was actually a VM [02:06] not off-hand; lots of random technical channels' -offtopic channels have knowledgeable people though [02:07] cwillu_at_work: nah. That level of access isn't really a concern of mine at the moment [02:07] jetole, in which case, access to the host is root to the vm :p [02:07] cwillu_at_work: I'd still like to find a means to manage these passwords if I can [02:08] I shouldn't be the only guy where I work to know the root and windows admin passwords [02:08] though it would be a nice level of job security [02:08] so don't be the only guy [02:08] give the boss an sealed envelop with the root password in it [02:08] cwillu_at_work: thats the point [02:08] don't make it complicated [02:09] cwillu_at_work: and if I choose to use a different root password on every host and we have a new host about twice a month? [02:09] jetole, I'll leave that as an exercise... :/ [02:09] hehe. Thanks [02:09] (hint: stationary is cheap) [02:10] well you have been helpful because you put auditing in a new light [02:10] I probably use paper more then anyone I work with but it is cumbersome in many applications [02:10] on another note [02:10] and speaking of auditing... [02:11] I live and die by my index cards :p [02:11] 4x6 <3 [02:11] (unruled) [02:11] what are some good means to audit actions on Linux servers? I was thinking perhaps prohibit shells through sudo (which is a one step configuration since we use sudo through LDAP) and have ldap email all commands [02:11] hmm, I should actually get some printed up as a business card come to think of it [02:12] jetole, that's a tough game to win [02:12] * jetole has a 6x8 notepad in front of him now and pretty much at all times when sitting at a desk [02:12] cwillu_at_work: yeah I figured at tough game to win [02:12] logging of all privilege-related stuff to a remote box is a start [02:13] I looked at, I think the package is auditd but it's a nightmare to decipher and still prone to manipulation [02:13] hello all i need help with updateing drivers for my Ethernet card but do not know the command to do this why this is, its cuz i have installed the server on my laptop then transferd it to my desctop since my desktop was unable to load the instalation [02:13] and then providing sufficient infrastructure that people don't feel the need to circumvent the rules to get work done [02:13] cwillu_at_work: logging to remote box how? [02:13] (if people use the hole they found every day because it's the only way to do foo, you've lost) [02:14] cwillu_at_work: well let's assume I trust everyone I work with. What are your thoughts on how to log this? [02:14] jetole, rsyslog and company can all write to remote daemons [02:14] for every command as the command is run? [02:14] no, but that's not really the point [02:14] oh [02:14] (I'm gonna have to run momentarily) [02:14] :-( [02:14] cwillu_at_work: thanks for the help [02:15] have you looked at what gets logged to /var/log/auth.log in the default config? [02:15] occasionally [02:15] * jetole takes a look now [02:15] you should probably have a good look :p [02:15] cwillu_at_work: not much [02:16] the important thing is that somebody doing something nefarious or illadvised can't do it without making at least one noteworthy log entry [02:16] cwillu_at_work: on my home 10.04 desktop I see a log of pam session opened stuff, some RSA stuff involving ssh(d) [02:16] gotta run [02:16] have fun [02:16] thanks again [02:16] gonna have to understand that stuff if you want a hope of implementing this :p [02:16] cwillu_at_work: well yeah. I was asking what it is I need to understand [02:17] jetole, rsyslog and related is a start ;p [02:18] cfc__: drivers are all kernel modules. You can view them with lsmod, remove them with rmmod, load them with modprobe and install a specific one that has the same name as a known modprobe one but in a different directory by using insmod [02:18] cwillu_at_work: again, thanks again [02:18] ;-) [02:18] cfc__: also have a look at /etc/modprobe.d [02:22] Could I get the assistance of an experienced Ubuntu user? To make a long story short Ubuntu crashed and now I cannot access my encrypted drive. Thank you. [02:23] diuneigh: why and how? [02:23] errors? console messages? etc? [02:23] •jetole• long story...1) installed playonlinux 2) it crashed while trying to get a game to work 3)when I rebooted it would not login and just loop back to the login screen. 4) I googled some solutions and tried them with a console Alt-f2 mode 5) after attempting to fix that problem I was unable to access my encrypted home dir. that is where I am now.. any ideas? [02:24] diuneigh: this is #ubuntu-server. Try joining the room #ubuntu [02:24] I already did. they suggested I try you as well. [02:24] diuneigh: your question sound very, very off-topic [02:24] diuneigh: then they were wrong [02:24] diuneigh: as the name of the room suggests, this is a room about ubuntu running on a server [02:25] diuneigh: I'm going to join #ubuntu now and see if I can help you there so let's follow up in that room but don't ask questions in the wrong room even if someone suggests you do [02:25] * jetole will continue this conversation in #ubuntu [02:25] •jetole• thanks... [02:26] •jetole• I am already in that room.. [02:47] stgraber: d'oh, lxc-delete should delete any symlink in /etc/lxc/auto/ [02:48] hallyn: it should indeed... [02:50] i'm just gonna file that as a low prio bug right now [02:51] smoser: ping [02:56] stgraber: ruh roh - useradd: group 'admin' does not exist [02:56] (creating an oneiric container) [02:56] oh, that's interesting ... [02:56] what's creating the admin group on oneiric then? [02:57] anybody know why http://x.x.x.x/cblr/svc/op/nopxe/system/foo would be failing to disable pxe boot on system foo? [02:57] when I wget that url (with my cobbler IP) I get back the content 'false' and netboot is still enabled [02:58] stgraber: no idea :) [03:01] hallyn: ok, will add a groupadd --system admin || true to the template [03:06] soudns good [03:07] stgraber: oh. so, you don't set root:root pwd at all any more? [03:07] right [03:07] didn't expect that :) probably for the best, though [03:08] the ubuntu way, one account with sudo rights [03:08] I guess that's what the cloud images do too [03:09] hallyn: empty lxc.devttydir != missing lxc.devttydir :( [03:09] hallyn: so the Oneiric container fails with: lxc-start: File exists - failed to create symlink for console [03:10] stgraber: all right then i'll need to update the patch. that's not for tonight [03:10] i'm trying to get the kernel patch out the door [03:12] hallyn: I have a patched template that will work with the current patch though (simply commenting the line) [03:15] hallyn: ok, the patched template works fine. I'll just try and make the default login/password more visible or I'm sure we'll get bug reports [03:15] so you'd want to push the package as is and send the patch fix later? [03:15] really it should be a one-line fix [03:17] stgraber: no wait, [03:17] stgraber: apparently we just need to use group 'sudo' everywehre? [03:18] I'm not completely sure how far back that worked and I know command-not-found and some others only started looking for 'sudo' in Precise [03:18] so we really should be using sudo only from Precise and use admin for previous releases [03:19] but admin doesn't exist in previous releases [03:20] so i end up with a lucid container where i can't sudo [03:20] no, the template I currently have will create the group if it doesn't exist [03:21] oh. sorry. misunderstood [03:21] all right, kernel patch away. SO isn't here glaring at me yet so lemme try see about fixing the patch fixed too for empty lxc.devttydir. if id isappear, i'll send it monday [03:28] stgraber: http://people.canonical.com/~serge/0040-consoles-into-devlxc.patch you just need the two new lines in confile.c from that patch [03:28] (though it also adds DEP3 (or is it DEP5?) tags) [03:29] hallyn: http://paste.ubuntu.com/837376/ [03:30] looks good [03:32] hallyn: http://paste.ubuntu.com/837379/ [03:32] hallyn: anything else or should I do a test build of that, confirm it works for oneiric and precise and then upload? [03:33] stgraber: unless you want to add the removng of /etc/lxc/auto/ symlinks on delete, i say push [03:34] (maybe test a lucid container too...) [03:34] btw the symlink removing could get murky... depending on how admins behave [03:39] hallyn: right, I guess it's not too bad in most cases because a devtmpfs is usually mounted on /dev, so even if they move their container back on a physical machine it should still work (or the case where you boot your VM in a container, then boot it back in kvm) [03:39] though it might be worth trying to cleanup on shutdown (but ignoring any error as we don't really know what the user does in the container) [03:46] New bug: #930525 in lxc (universe) "lxc-destroy should remove symlink from /etc/lxc/auto/ if it exists" [Undecided,New] https://launchpad.net/bugs/930525 [03:47] smoser ? [03:47] smoser what was the issue with thqt half installed package [03:54] stgraber: oh, sorry, i meant just the removing of the /etc/lxc/auto symlinks. though yeah, the /dev/console symlink might bite us one day [03:55] hallyn: oh right, different symlink ;) [04:17] hallyn: uploading lxc now [04:17] hallyn: confirmed to work fine on Oneiric, can't test precise as debootstrap fails (worked fine with an existing container) [04:28] hallyn: uploaded [07:36] New bug: #930551 in quota (main) "System with quota enabled hangs after upgrade to 11.10" [Undecided,New] https://launchpad.net/bugs/930551 === CasmoNL_ is now known as CasmoNL [11:37] Hello I want to ask if anyone kan help me set up my apache server so that it connects to the dir public_html. Have tried to read a few guides but it does not seem to work. [11:49] so a per user share or is it just you? [11:53] have you tried a2enmod mod_userdir ? [11:58] that should be sudo a2enmod userdir [11:58] yes as a per user share. Have tried that. I am now looking up the file that I used. [12:01] well first you need the apache module enabled, then put the following in /etc/apache2/sites-avalable/default :- UserDir public_html [12:06] is it correct with the following to enable the module, [12:08] I edit this file: /etc/apache2/apache2.conf [12:09] I add the following lines, [12:09] # Turn on userdirs [12:09] [12:09] UserDir public_html [12:09] [12:09] Order allow,deny [12:09] Allow from all [12:10] [12:10] [12:10] # [12:11] is that correct? [12:24] are u there? [12:25] sorry i was distracted [12:25] !pastebin [12:25] For posting multi-line texts into the channel, please use http://paste.ubuntu.com | To post !screenshots use http://imagebin.org/?page=add | !pastebinit to paste directly from command line | Make sure you give us the URL for your paste - see also the channel topic. [12:26] so if the module is enabled use that rule, is the module enabled? use a2enmod userdir is the "debian way" [12:27] clearwood: debian/ubuntu uses a set of symlinks in /etc/apache2/mods-enabled to enable/configure mods - the way you use here, is the hard core apache way, which isn't the 'normal' debuntu way [12:27] normally you would put that in a vhost config ie /etc/apache2/sites-avalable/default (or another file in that dir) which you have also run a2ensite filename [12:28] dns53: erm - no [12:28] dns53: the userdir is normally global [12:28] it might be good to use it under a vhost, though, but that's a choice to make [12:29] RoyK ok, i have not used it myself but i know it can be done [12:29] where do I use a2enmod userdir in the file apache2.conf ? [12:29] I use userdir a bit... [12:29] clearwood: a2enmod is a command [12:29] not a configuration setting [12:30] a2enmod creates a symlink under /etc/apache2/mods-enabled pointing to the module and its config under /etc/apache2/mods-available [12:30] creating the symlink manually does the same thing [12:30] but then I have to type it every time I turn the server on dont I [12:30] last I checked, symlinks were persistent ;) [12:31] ok Ill try it thanks [12:31] remember to restart/reload apache after that [12:31] apache2ctl restart/graceful/something [12:31] yes [12:32] I use this : sudo /etc/init.d/apache2 restart [12:32] that works too [12:34] I then get this response: Module userdir already enabled [12:35] so my problem must be that the directory public_html is not connected [12:36] I allso put the following in /etc/apache2/sites-avalable/default UserDir public_html [12:36] and rstarted [12:37] upon restart I get ... waiting apache2: Could not reliably determine the server's fully qualified domain name, using 127.0.1.1 for ServerName [12:38] when I try to access through firefox with [12:38] http://127.0.1.1/ I get [12:39] It works! [12:39] This is the default web page for this server. [12:39] The web server software is running but no content has been added, yet. [12:39] and I know that theres something in my public_html directory [12:39] but it does not show up?? [12:40] so user directories would be http://127.0.0.1/~username/ [12:40] Then I get the following: [12:40] Forbidden [12:40] You don't have permission to access /~rub on this server. [12:41] ls -l|grep public_html [12:42] gives me: [12:42] drwxrwxr-x 3 rub rub 4096 2012-01-28 09:59 public_html [12:42] is that by the way to unsecure? [12:44] you need write on a directory to be able to look inside [12:44] unless another user is in the "rub" user group you should be fine [12:45] have you created a .htacces file and included a config line that denied access to the directory? [12:49] could you try removing your own config from your apache2 config, i wonder if it is using the /etc/apache2/mods-available/userdir.conf file as well as your own config [13:28] I lost my connection I am back [13:29] where do I find the former comments are they lost forever? [13:31] fighting with tryin to set up public_html to my apache server [13:32] used a2enmod userdir [13:33] but no connection to the directory with firefox to http://127.0.1.1/~username [13:34] does anyone know how to set up acces? [13:34] please :) [13:37] is any one there? [13:40] clearwood: former comments are at http://irclogs.ubuntu.com/2012/02/11/%23ubuntu-server.html [13:53] Why does my whole '/' seem to be read-only? === med_out is now known as med_ [14:21] New bug: #930652 in lxc (universe) "lxc & swapoff: Not superuser" [Undecided,New] https://launchpad.net/bugs/930652 === medberry is now known as med_ [16:15] Hello! I need some help with NFS configuration. I'm not able to access filesystem on server from client with the following settings - http://paste.ubuntu.com/837967/ Is the config right? How can I fix it? [16:19] da what? [16:19] you can't export disks [16:19] you have to export folders [16:20] Its filesystem what matters isn't it? [16:20] Well in the end of the paste, you can see the config which worked before [16:20] that doesn't invole stuff in /dev [16:21] yes [16:21] nfs can't handle filesystems [16:21] it doesn't work on block level devices [16:21] use nbd if you want that [16:21] or the ataoe [16:22] Alright. Is there any command to run after changing exports file on server? [16:22] yes [16:22] been too long, don't remember [16:22] restart nfsd and portmap process? [16:23] ah-ok [16:24] What do you use for such purpose? [16:43] orated: to restart services? [16:44] service foo restart|start|status [16:45] sudo /etc/init.d/portmap restart and same for restarting nfs-kernel worked [16:46] s/same/likewise* [16:53] init.d scripts are the pre-upstart method of running services, some things still provide init.d scripts but most common services now use upstart scripts [16:53] anyhow, init.d scripts or upstarts scripts will work. [16:54] given the security issues of NFS, its bad a idea to export your root directory to NFS [16:57] ah.. yes. Thanks :) BTW I was trying to edit exports. / 192.168.10.0/24(rw,no_root_squash,async) and /home 192.168.10.0/24(rw,no_root_squash,async) in exports on server and then 192.168.10.101:/ /mnt/ nfs rsize=8192,wsize=8192,timeo=14,intr 0 0 & 192.168.10.101:/home /mnt/home nfs rsize=8192,wsize=8192,timeo=14,intr 0 0 in fstab on client gives - mount.nfs: Stale NFS file handle error twice [17:02] make use previous mount attempts have been umount'd and try to mount to somewhere else [17:03] /mnt/nfs1 for example [17:03] Yes, tried that - umount.nfs: /mnt: device is busy [17:04] sec [17:04] force with `umount -f` [17:05] I got same error with - umount2: Device or resource busy [17:08] hmm, perhaps try `umount -l` for 'lazy umount' [17:08] unilateral umount is more accurate description imo [17:09] great! [17:10] lol, i've just realised i'm #ubuntu-server not #ubuntu-uk [17:11] heh [17:17] hi,i am trying to find out why my page speeds are so slow, 10.53761 seconds the server speed is very fast using this online tool http://www.serverscores.com/speedtest [17:17] default vhost on apache [17:18] what test gives you 10.5 seconds? [17:22] i'd guess that 'server speed' is ping and page speed is to receive the http url [17:22] yes [17:22] why would it be so slow [17:22] does your page depend on any database queues? [17:23] its default apache, [17:23] 'It works' page [17:24] i would like to say its DNS, but thats from server IP [17:24] slow upstream? [17:24] the server is lightning [17:25] DNS is common to ping and http retreive and would be cached anyway. [17:25] lightning? [17:25] get 30mb/s from it [17:25] download [17:26] what can effect page load? [17:26] i mean, its defautl apache2 vhost, what about PHP settings? they are also default [17:27] the php5 module isn't enabled by default in ubuntu server [17:27] obviously i have installed it.. [17:27] anyhow, the 'it works' page doesn't invoke the php interpretter [17:27] neither is apache2 [17:28] ok good point. [17:28] how else can i test this? [17:28] locally ? [17:28] that's true, but one might expect the apache2 package to include and enable php5 [17:28] you need to find out where it is slow from, i.e. at point between you and the server does it become slow? [17:29] how can i test page speed load locally? [17:29] everything else seems fine [17:30] i will PM you the IP to show how slow it is? [17:30] jacobw: I tried unmounting and remounting to reset NFS.. What used to mount before doesn't mount now and vice-versa [17:31] Tixos: try httperf [17:32] ok, can you load the page for me and report what its like ? [17:32] Tixos: how long ago was this server brought online? [17:33] hm, [17:33] 3days 30minutes [17:35] is this a bad result? > Reply time [ms]: response 9328.6 transfer 0.0 [17:35] from what? [17:35] not sure how to read this output (httpref) [17:35] should i pastebin ? [17:36] http://pastebin.com/FRNhkPhn [17:37] Tixos: i doubt this is a problem with apache [17:38] do those results look normal? why would you conclude that ? [17:41] Tixos: there's most likely nothing wrong unless can you prove a slow connection to upstream [17:42] you mean like this? [17:42] httperf --server=www.google.com [17:42] Request rate: 4.1 req/s (244.0 ms/req) [17:42] my server [17:42] Request rate: 0.1 req/s (18697.6 ms/req) [17:43] Tixos: how is your server provided? [17:43] how do you mean ? its a dedi [17:45] orated: stop nfs on the server and try to umount those points again [17:45] jacobw: Yes, on it [17:46] Tixos: how much bandwidth should you have? [17:47] unlimited [17:47] jacob, these results dont indicate something wrong with http? [17:48] they indicate that http to your server is slow from where you are right now [17:48] can you help me debug man :( [17:49] im putting my traffic onto another server for the minute, to see if it effects the server [17:49] ealier before i swtiched to this server, it was so fast [17:49] fast http? [17:49] jacobw: / and /home access from client works good. But not the /media.. [17:49] never used httperf, but yes load time view browser was less than 1 second [17:50] i have no firewall etc, i cant think of anything else! [17:50] other server is centos, and has no issues, is there any default ubuntu apache2 configs that could cause this if the server receives alot of traffic? [17:51] the apache configurations should be similar, both should use virtual hosts by default [17:52] yes, what about client limit ? i read somewhere [17:52] im out of ideas, so i came ehre :( [17:52] just moved traffic to backup server [17:52] where is the other server in relation the ubuntu server? [17:52] so i will give it 5minutes, and check again with httperf, like i said this morning it was very quick [17:52] same country [17:53] same network? [17:53] not sure, do you think it could be beyond the dedi then ? [17:53] like a global IDS for the rack or something? [17:54] ping is fast enough [17:55] yea i know [17:55] GRR [17:55] httperf tells us that http is slow [17:55] so why would it be? [17:55] config? [17:56] got slower [17:56] if the centos host is on the same network, compare apache configurations, if not use iperf to test tcp speed from your location to the ubuntu server [17:56] faster [17:56] Request rate: 0.1 req/s (9849.8 ms/req) [17:59] if the centos and ubuntu servers are on the same network, you can test between them to rule out the ubuntu server as the cause [17:59] they are not [18:01] ok, use iperf to test tcp speed between you and the server, if that's good then you know its the http server, otherwise contact the provider and ask for a test nearer the server, i.e. from before upstream [18:04] how to use iperf? >$ iperf -s IP ? [18:07] http://rackerhacker.com/2010/03/20/testing-network-throughput-with-iperf/ [18:10] hey jacobw! Its working fine now after restarting almost everything [18:11] :) [18:11] Thanks [18:12] But I think I can have problems with NFS. Could you tell me how NBD does the same job is a better manner? [18:12] it doesn't really [18:14] actually, i'm not so sure, it might work well if your on a fast LAN [18:14]