| dustinspringman | thank you pmatulis, I'm reading into it now. | 00:04 |
|---|---|---|
| dustinspringman | it appears, though I probably wont know for sure for a bit, that doing apt-get update and upgrade have resolved the issue.. pppd is no longer stacking up and all my clients are connecting and I'm only using 1% of my CPU now... | 00:20 |
| pmatulis | dustinspringman: sysstat package is great for logging a history of used resources. dstat package is good for realtime | 00:27 |
| dustinspringman | pmatulis: thanks for that, I'll give those a shot. so far no pppd creep! | 00:27 |
| === Lcawte is now known as Lcawte|Away | ||
| cyclob|work | hi guys, how do i stop libreoffice 4.2 from autostarting on boot on 14.04? | 00:55 |
| pmatulis | cyclob|work: how is that a ubuntu server question? | 00:56 |
| cyclob|work | cause i'm running ubuntu server? | 00:56 |
| TJ- | cyclob|work: See the channel topic, "For general (not server specific) support, try #ubuntu" | 00:58 |
| dustinspringman | pmatulis: it came back.. but I was able to use sysstat and dstat to find the issue! | 01:05 |
| pmatulis | dustinspringman: oh nice. what was it? | 01:06 |
| dustinspringman | pmatulis: pptp was using a plugin to log client connection/disconnections.. but that was eating CPU like cray cray | 01:06 |
| pmatulis | dustinspringman: i see, you disabled it then? | 01:06 |
| dustinspringman | pmatulis: I commented out the logging plugin from /etc/ppp/pptpd.conf and boom. 3% avg CPU load as expected. | 01:07 |
| pmatulis | dustinspringman: sweet, nicely done | 01:07 |
| dustinspringman | pmatulis: thanks for the tip on those tools, those are mucho helpful | 01:08 |
| jezeniel | Any upstart users active here? | 01:10 |
| pmatulis | jezeniel: i suppose we're all upstart users here, was'up? | 01:10 |
| dustinspringman | jezeniel: I've been an upstart for nearly a decade! XD | 01:11 |
| jezeniel | I just have some noob questions.. Can upstart have multiple "start on" and "stop on" clauses? | 01:12 |
| dustinspringman | jezeniel: I'm unfamiliar with "upstart" are you talking about a specific package or is this the "boot process" you are asking about? | 01:18 |
| firesword13 | So, I'm having problems getting phpbb3 set up on 12.04. | 01:33 |
| firesword13 | I was following the guide located at https://help.ubuntu.com/community/PhpBB3 | 01:34 |
| firesword13 | But when I went to start the part in the browser, I get a general error saying access is denied. | 01:36 |
| firesword13 | The threads I've found are all along the lines of "it used to work, but now it doesn't" whereas mine hasn't worked from the start. | 01:37 |
| phuh | What do you call these things in BASH? ${var:=xxx} ${var:-xxx} | 01:40 |
| firesword13 | So no one has any suggestions about my phpBB3 problem? | 02:06 |
| firesword13 | Well, I suppose I'll try asking again tomorrow. | 02:35 |
| === edsoncanto is now known as edson | ||
| === edson is now known as edsoncanto | ||
| dustinspringman | okay, thought I had this fixed, but now I an CERTAIN I know what the cause is.... when pptp clients are disconnecting (for whatever reason, gracefully leaving, service interruption, etc) my servers CPU gets a task listed in top that will eat 100% of the CPU if not killed.... obviously this is a problem! Any thoughts on how to stop pppd from eating the CPU when a client disconnects? | 03:41 |
| === Guest7277 is now known as foo | ||
| === Ursinha is now known as Ursinha-afk | ||
| lordievader | Good morning. | 06:37 |
| === vincent_1dk is now known as vincentvdk | ||
| Vladimir_ | I need to append time/date to a file name when it arrives to a directory from a tftp client, do I have to write a bash script or what do I use? | 08:49 |
| lordievader | Vladimir_: What do you use tftp for? | 09:00 |
| Vladimir_ | lordievader: its either tftp or sftp | 09:03 |
| Vladimir_ | but it doesn't matter, I just need to send files locally to a linux machine | 09:03 |
| lordievader | Vladimir_: Then use sftp, I thought tftp isn't very efficient. But that might be me. | 09:09 |
| lordievader | Vladimir_: But if the ctime does not satisfy, then you likely need to write a script that checks a folder. Throws some regex on the filenames and if they don't follow the rules change the name. | 09:10 |
| Vladimir_ | lordievader: it doesn't make any difference for me, all I need it to append date/time to the files | 09:10 |
| Vladimir_ | lordievader: alright=) | 09:11 |
| Vladimir_ | lordievader: do you know if it's possible to set that "If the file comes from a specific IP, THEN add the hostname(or a specific word)" ? | 09:12 |
| lordievader | Vladimir_: The protocol doesn't do that, however you might find a server capable of doing so. | 09:14 |
| Vladimir_ | lordievader: okey, Oh so UDP doesn't send the source IP ? | 09:15 |
| lordievader | Vladimir_: It does, but SFTP is only a transfer protocol. It doesn't really do management, that is left up to the server/administrator. | 09:16 |
| Vladimir_ | lordievader: ok, can the linux machine see if a specific file did come from a IP adress? | 09:19 |
| Vladimir_ | otherwise I can just set in my expect script the name of the backupfile to the hostname and then append only time/date on the linux machine(tftp server) | 09:20 |
| lordievader | Vladimir_: If one user per host is used it can be tracked that way. | 09:23 |
| Vladimir_ | lordievader: you mean that there's only one user on the client(switch device in my case) ? | 09:25 |
| lordievader | No, user1 maps to client1, user2 maps to client2, etc. | 09:26 |
| Vladimir_ | hmm ok | 09:26 |
| sacarde | hi | 09:43 |
| sacarde | do you know how remove the "clear" before login | 09:45 |
| sacarde | in tty1 | 09:45 |
| Vladimir_ | lordievader: Do I have to use any programming och bash scripting language to append timestamp to all the files coming in to a certain directory? | 09:47 |
| lordievader | sacarde: http://askubuntu.com/questions/58097/how-can-i-remove-the-clear-screen-before-login | 09:48 |
| lordievader | Vladimir_: Not if you take my method of regex. For <-- that bash will do fine. | 09:48 |
| Vladimir_ | okey | 09:49 |
| sacarde | lordievader, ok, this is | 09:50 |
| sacarde | thanks a lot | 09:50 |
| ertyos | hello there | 09:51 |
| ertyos | i got sendmail on existing server | 09:52 |
| ertyos | i simply create mail account using this tutorial : http://paste.ubuntu.com/8214092/ | 09:54 |
| Vladimir_ | lordievader: so I can use regex instead of ctime right? | 09:54 |
| ertyos | but the problem is the account i created not working with mail agent (thnunderbirrd) | 09:55 |
| ertyos | what to do ? | 09:55 |
| lordievader | Vladimir_: Err, what I ment was that you check with a few regex if a filename complies with the rules you've set. If it doesn't you rename it (based on the ctime, or something) so it does comply with the set rules. | 09:56 |
| === Lcawte|Away is now known as Lcawte | ||
| === MikaT_ is now known as MikaT | ||
| === shauno_ is now known as shauno | ||
| Vladimir_ | lordievader: lordievader: ok, but the only thing I need is a script that checks if a folder has any new files, and if they have then add the current date to the filename | 10:54 |
| lordievader | Vladimir_: Read my answer again, it does relatively the same. | 10:56 |
| Vladimir_ | lordievader: sorry, i'm logged in on webchat.freenode, it tends to log me out every 30 minuter :/ | 10:57 |
| pmatulis | morning | 11:00 |
| === sync0new is now known as sync0pate | ||
| lordievader | Vladimir_: Err, what I ment was that you check with a few regex if a filename complies with the rules you've set. If it doesn't you rename it (based on the ctime, or something) so it does comply with the set rules. | 11:03 |
| lordievader | Hey pmatulis, how are you? | 11:03 |
| dustinspringman | pmatulis: you still here? | 11:29 |
| pmatulis | dustinspringman: yup | 11:48 |
| dustinspringman | pmatulis: I've got to run for now, but I've isolated the cause of the rouge pppd 100% cpu usage issue... | 12:21 |
| dustinspringman | pmatulis: when a pptp tunnel disconnects, the pppd process starts and just eats and eats CPU... I have no idea what to do to get that to stop! I thought it was the logging plugin in the /etc/ppp/pptp.conf, but that only stopped "logged in" logs from hogging CPU.. this disconnect/log process, i've been unable to find.. =/ | 12:23 |
| patdk-wk | don't use pptp? | 12:23 |
| patdk-wk | it's been over a decade since I know anyone that used pptp | 12:24 |
| dustinspringman | patdk-wk: what protocol are you recommending? | 12:24 |
| patdk-wk | ipsec? | 12:24 |
| dustinspringman | patdk-wk: many of the routers I have to connect to this vpn aggregation point are not ipsec compatible.. =/ | 12:25 |
| patdk-wk | something that isn't ipsec compatable exists? | 12:25 |
| patdk-wk | that I just find amazing | 12:25 |
| dustinspringman | **requires additional licenses the customer doesn't care to pay for | 12:25 |
| patdk-wk | what brand? I'll make sure I stay clear | 12:26 |
| patdk-wk | but I normally use cisco, and it's included by default there | 12:26 |
| dustinspringman | sonicwall.. they got the super low budget ones... some other guy engineered the networks, I just took them over.. | 12:26 |
| dustinspringman | pptp works fine for this purpose, its just to connect a central monitoring server to watch all the remote sites... but the server CPU spiking when a client disconnects is problematic.. | 12:27 |
| dustinspringman | gotta run, be back later. | 12:29 |
| RoyK | dustinspringman: openvpn? | 12:45 |
| patdk-wk | on sonicwall appliance? | 12:45 |
| RoyK | probably not :P | 12:45 |
| === mjohnson151 is now known as mjohnson15 | ||
| lordievader | Wasn't pptp broken? (As in insecure) | 12:50 |
| === smoser` is now known as smoser | ||
| eutheria | so i thought i would setup ipv6 | 13:23 |
| eutheria | i use dhcp to configure hosts on the network, i don't give static ip addresses | 13:23 |
| eutheria | so openwrt and it seems to be assigning ipv6 addresses to desktop clients, but my server doesn't seems to be picking up an ipv6 address | 13:24 |
| eutheria | do i need to enable something in the interfaces file to enable ipv6 ? i can't see anything in the man page unless i am being dumb | 13:25 |
| eutheria | iface eth0 inet6 dhcp maybe | 13:26 |
| Vladimir_ | hello | 13:27 |
| Vladimir_ | Is it easy to change the comments color in Vim/Vi ? | 13:28 |
| Vladimir_ | I have a very irritating blue colour which I can't even se because of the black background | 13:28 |
| === dannf` is now known as dannf | ||
| === alai` is now known as alai | ||
| === deegee is now known as drussell | ||
| jamespage | zul, this will make neutron vpn a little tricky | 13:53 |
| jamespage | https://launchpad.net/ubuntu/+source/openswan/+publishinghistory | 13:53 |
| === Ursinha-afk is now known as Ursinha | ||
| jamespage | jpds_, hows your strongswan stuff coming along upstream? | 13:54 |
| jpds_ | jamespage: What bits, where? | 13:55 |
| jamespage | jpds_, I thought that you submitted some code upstream in openstack for strongswam support in neutron? | 13:56 |
| jamespage | maybe I was wrong | 13:56 |
| jpds_ | jamespage: I did, and it's going into Juno-3. | 13:56 |
| jpds_ | jamespage: Last I heard. | 13:56 |
| jamespage | jpds_, awesome | 13:56 |
| jamespage | jpds_, openswan got removed from utopic so we need an alternative | 13:57 |
| jpds_ | jamespage: Luckily most of openswan as nothing to do with strongswan. | 13:57 |
| jpds_ | jamespage: There is a problem with the precise package of strongswan which I'm working on. | 14:01 |
| eutheria | weird iface eth0 inet6 dhcp brings up the ipv4 address | 14:02 |
| ertyos | hello | 14:11 |
| ertyos | i got an sendmail user which is not present on /etc/passwd is it normal ? | 14:12 |
| ertyos | i got sendmail working user which is not present on /etc/passwd | 14:12 |
| ertyos | is it normailL ? | 14:12 |
| ertyos | no idea ? | 14:13 |
| jpds_ | !repeat | 14:14 |
| ubottu | Don't feel ignored and repeat your question quickly; if nobody knows your answer, nobody will answer you. While you wait, try searching https://help.ubuntu.com or http://ubuntuforums.org or http://askubuntu.com/ | 14:14 |
| bananapie | Can I boot to a USB key from grub on a hard disk on a computer whose BIOS doesn't support booting to USB ? | 14:21 |
| ertyos | any one there ? | 14:22 |
| jpds_ | ertyos: Yep. | 14:22 |
| jpds_ | ertyos: Personally, I don't use sendmail, I think most people use postfix. | 14:22 |
| cfhowlett | bananapie, nope. | 14:23 |
| bananapie | bummer. | 14:23 |
| ertyos | don't care | 14:23 |
| ertyos | wheter postfix or sendmail | 14:23 |
| bananapie | My media centre ( mythbuntu ) is on an older p4 that doesn't boot from USB sticks. I want to reinstall it with 14.04 and I don't want to burn a CD. :( | 14:23 |
| ertyos | it is someoff system related question | 14:23 |
| zul | jamespage: where is the nova-comptue-vmware charm? | 14:23 |
| jpds_ | jamespage: zhhuabj wrote the patch for strongSwan. | 14:23 |
| jamespage | zul, in the charm store | 14:24 |
| zul | jamespage, right | 14:24 |
| eutheria | interesting dhclient doesn't register a hostname by default with ipv6 | 14:24 |
| zhhuabj | jpds, jamespage: yeah, I wrote it | 14:24 |
| ertyos | when you create user, when the user is working correctly with your mail agent and not appearing on the file /etc/passwd and can't change that user using passwd user is it that normal or not ? | 14:24 |
| bananapie | How often do you guys reinstall your production servers or upgrade to the next LTS ? | 14:25 |
| eutheria | bananapie: i have a 12.04 server i have no plans to upgrade anytime soon | 14:26 |
| eutheria | it is really down to what works for you how often you upgrade, what risks etc | 14:27 |
| jamespage | zhhuabj, awesome - thankyou! | 14:28 |
| zhhuabj | jamespage: K opens for spec proposals is on Sep 4 based on https://wiki.openstack.org/wiki/NeutronJunoProjectPlan | 14:29 |
| === cmagina_ is now known as cmagina_airport | ||
| bananapie | I found this thread on reddit, http://www.reddit.com/r/linux/comments/2f3dih/please_help_me_update_security_updates_on/ck5whbl | 14:31 |
| === cmagina_airport is now known as cmagina_mobile | ||
| bananapie | Have I misunderstood this guy's post? | 14:31 |
| jamespage | zhhuabj, do you know whether strongswan support will land for juno? | 14:31 |
| zhhuabj | jamespage, jpds, we need to move strongswan neutron-spec (https://review.openstack.org/#/c/101457/) from Juno to K after Sep 4, and push some core approver to review it and let it merge | 14:32 |
| jpds_ | zhhuabj: Problem is that leaves us VPNless for Juno. | 14:33 |
| zhhuabj | jamepage: no, will not land for juno because Juno's spec approval deadline (SAD) is Jul 20 based on https://wiki.openstack.org/wiki/NeutronJunoProjectPlan | 14:33 |
| zhhuabj | jamespage, jpds, for the neutron-spec https://review.openstack.org/#/c/101457/, the neutron PTL Kyle give one +2 on Jul 15, but no another core approver to merge the spec after SAD (Jul 20) has passed, so it can't catch up Juno | 14:38 |
| zul | hallyn: ill do libvirt 1.2.8 this afternoon | 15:13 |
| caribou | jamespage: FYI, I won't be able to attend the meeting this week | 15:33 |
| caribou | jamespage: nothing particular to bring up though | 15:33 |
| jamespage | caribou, ack | 15:33 |
| eagles0513875 | hey guys | 16:00 |
| eagles0513875 | i just upgraded to 12.04.1 and i am getting http://pastebin.com/9RYNWJh8 | 16:01 |
| eagles0513875 | my httpd.conf file has disappeared | 16:01 |
| eagles0513875 | any ideas why | 16:01 |
| lordievader | eagles0513875: Apache 2.2 -> 2.4? Also 12.04.1? Do you mean 12.04.5? | 16:04 |
| eagles0513875 | i have one server on 12.04.1 which upgraded to 2.4 and another on .5 and still on 2.2 | 16:04 |
| eagles0513875 | O_o | 16:04 |
| eagles0513875 | should i purge apache and set it up again | 16:04 |
| eagles0513875 | and if i purge it would i lose my vhosts? | 16:04 |
| lordievader | eagles0513875: Rather check where those packages are coming from. | 16:05 |
| eagles0513875 | not sure if a sudo do-release upgrade to .1 would make a difference | 16:05 |
| ikonia | no | 16:05 |
| ikonia | you've installed PPa's again | 16:05 |
| ikonia | release upgrade won't take you to .1 | 16:06 |
| ikonia | 12.04 ships with apache 2.2 | 16:06 |
| ikonia | so you have a 3rd part repo pulling in 2.4 from soewhere | 16:06 |
| ikonia | somewhere | 16:06 |
| eagles0513875 | ok seems like i got upgraded to 2.4 | 16:06 |
| eagles0513875 | i mean to 14.04 | 16:06 |
| ikonia | yes, because you did do-release-upgrade | 16:06 |
| eagles0513875 | ok now to fix the issues | 16:07 |
| eagles0513875 | what can i do about apache with its missing httpd.conf file | 16:07 |
| ikonia | what do you mean missing ? | 16:07 |
| lordievader | eagles0513875: Read the documents about the differences between 2.2 and 2.4. | 16:07 |
| eagles0513875 | lordievader: the release notes? | 16:07 |
| ikonia | we've had the discussion on 2.2/2.4 differences about 10 times now, you've also had them explained to you in #httpd | 16:08 |
| lordievader | eagles0513875: Google: apache 2.2 vs 2.4 and you'll find all about it. | 16:08 |
| eagles0513875 | im aware of the vhost differences | 16:08 |
| eagles0513875 | but when i try and start apache i get the error apache2.conf is trying to find httpd.conf for some reason which cant be correct | 16:09 |
| ikonia | why can't that be correct ? | 16:09 |
| ikonia | why is the error message it's giving you "not correct" | 16:09 |
| eagles0513875 | ikonia: what im not understanding is why would it look for that file if it has been removed | 16:10 |
| ikonia | who said it's been removed ? | 16:10 |
| ikonia | who said this is the stock config | 16:10 |
| ikonia | you don't really appear to be %100 sure of what you are doing with your machine, so who knows what you've done | 16:10 |
| ikonia | eagles0513875: try actually looking at the config and see what / where it's including that file | 16:10 |
| eagles0513875 | ok | 16:10 |
| ikonia | instead of saying "that error is wrong" fix the error | 16:10 |
| ikonia | accept the error is correct and resolve the issue | 16:10 |
| eagles0513875 | there is # Include all the user configurations: Include httpd.conf should i comment that out? | 16:12 |
| ikonia | eagles0513875: depends, | 16:16 |
| ikonia | apply thought | 16:16 |
| eagles0513875 | ok | 16:16 |
| lordievader | eagles0513875: With common-sense and a bit of Google you can go far ;) | 16:27 |
| === utlemming is now known as utlemming_away | ||
| xibalba_ | hey folks, have a few newb questions for you. whats the best, in terms of docs/compatibility/etc..., virtualization engine to use on ubuntu-server? kvm/xen? | 17:38 |
| xibalba_ | i guess whats supported right out of the box | 17:38 |
| xibalba_ | i just need 1 single physical box, running a few smaller ubuntu boxes as vms | 17:38 |
| pmatulis | xibalba_: kvm | 17:40 |
| xibalba_ | ok, any recommended intro docs for it? i come from a heavy vmware background | 17:40 |
| xibalba_ | this looks like a good start, https://help.ubuntu.com/community/KVM | 17:40 |
| pmatulis | xibalba_: yes, plus the ubuntu server guide | 17:40 |
| xibalba_ | anything in particular out of the server guide? i've been running unbuntu server vms for a while, just never layered on virtualization | 17:41 |
| pmatulis | xibalba_: yes, the virtualization section. LXC is also good | 17:42 |
| xibalba_ | LXC is similar to Docker, correct? | 17:42 |
| xibalba_ | lxc > libvirt ? | 17:44 |
| patdk-wk | !best | 17:45 |
| patdk-wk | !poll | 17:45 |
| patdk-wk | damned bot | 17:45 |
| xibalba_ | ha | 17:46 |
| patdk-wk | well, libvirt is generic | 17:46 |
| patdk-wk | kvm is simular to esxi | 17:47 |
| patdk-wk | lxc is like openvz | 17:47 |
| patdk-wk | docker is build around lxc, I believe | 17:47 |
| patdk-wk | libvirt is a generic lib, to managethem all, basically | 17:47 |
| xibalba_ | i think kvm will work best for me since its the closest thing i'm use to for now | 17:48 |
| xibalba_ | i definetly have plans to check out container based virtualization | 17:49 |
| patdk-wk | I used to use it | 17:49 |
| patdk-wk | gave it all up, cause of management issues | 17:49 |
| patdk-wk | just easier to update and upgrade full vm's for me, than containers | 17:49 |
| xibalba_ | understood | 17:49 |
| patdk-wk | lost about 15% performance, but heh | 17:50 |
| === mikal_ is now known as mikal | ||
| xibalba_ | yeah not super concerned on the performance side since it's just a personal box | 17:50 |
| xibalba_ | 15% lost on containers though? | 17:50 |
| xibalba_ | thats pretty significant | 17:50 |
| sarnold | xibalba_: I think patdk-wk meant using VMs was a 15% performance loss compared to containers.. | 17:51 |
| patdk-wk | yes, loss using vmware, vs using containers | 17:51 |
| patdk-wk | but that was before my cpu's where ept enabled | 17:52 |
| patdk-wk | likely only 10% loss now | 17:52 |
| xibalba_ | ah ok | 17:52 |
| xibalba_ | yeah our procs at work at so beefy its neglible | 17:52 |
| patdk-wk | heh, I have new 24core dual socket systems | 17:53 |
| patdk-wk | they are not beefy enough | 17:53 |
| patdk-wk | 24core 2.9ghz I think they are | 17:53 |
| patdk-wk | ah, 2.7ghz | 17:54 |
| sarnold | very nice | 17:54 |
| patdk-wk | ms sql with 3tb db (spanned over 48 ssd's) still ownes the box | 17:55 |
| xibalba_ | that is a beeeeeefy database | 17:55 |
| xibalba_ | whats running on it | 17:55 |
| patdk-wk | everything the batch job gets queued up, vmware shifts all load from that box, to the others :) | 17:55 |
| patdk-wk | just batch processing | 17:56 |
| patdk-wk | just recomputing all data links every month | 17:56 |
| patdk-wk | as more infomation comes in, or gets more accurate | 17:56 |
| sarnold | ahhh, that's why 48 ssds rather than fusionio, heh? | 17:56 |
| patdk-wk | so we can't depend on the last links to be accurate or valid | 17:56 |
| patdk-wk | as they where only the best guess at that time | 17:57 |
| patdk-wk | we looked at fusionio | 17:57 |
| patdk-wk | just didn't pan out as a workable solution | 17:57 |
| patdk-wk | this is actually a pure system | 17:57 |
| patdk-wk | so the san ssd backs all the esxi hosts | 17:57 |
| xibalba_ | are you using RDMs for that DB ? | 17:58 |
| patdk-wk | no | 17:58 |
| xibalba_ | is your DB spread across multiple VMDKs then? | 17:58 |
| patdk-wk | ya, like 10 :) | 17:59 |
| xibalba_ | also do you prefer NFS or iSCSI for mounting your SAN to your ESXI boxes? and why | 17:59 |
| patdk-wk | well, considering, nfs sucks for that workcase? | 17:59 |
| xibalba_ | other than, my SAN doesn't support NFS thats what a NAS is for =P | 17:59 |
| patdk-wk | nfs is limited to a single ip connection | 17:59 |
| xibalba_ | does it? | 17:59 |
| patdk-wk | so you want to limit it to 10gbit? | 17:59 |
| xibalba_ | ah so you're using multipathing? | 17:59 |
| patdk-wk | well, 4 fc channels | 18:00 |
| xibalba_ | ok, we took a slightly different approach to our NFS mounts. By mounting the volumes from the NAS on off-by-one IPs and using IP HASH as the port-channel alogirthm on the switches | 18:00 |
| xibalba_ | so 192.168.1.220:/vol1 | 18:00 |
| xibalba_ | 192.168.1.221:/vol2 | 18:00 |
| xibalba_ | etc | 18:00 |
| patdk-wk | yes, but your still limited to a single nic port per nfs mount | 18:01 |
| patdk-wk | there is no way to fix that | 18:01 |
| xibalba_ | correct, not till pnfs4 | 18:01 |
| xibalba_ | http://www.pnfs.com/ | 18:01 |
| patdk-wk | Ialso dislike the additional overhead of svmotion on nfs | 18:01 |
| xibalba_ | oh w/out VAAI support you mean? | 18:01 |
| patdk-wk | you can have vaai support on nfs? | 18:02 |
| xibalba_ | i believe so, on the netapps | 18:02 |
| patdk-wk | on nfs? I understand iscsi | 18:02 |
| xibalba_ | https://library.netapp.com/ecmdocs/ECMP1237939/html/html/GUID-735E5961-E3FB-4105-A8F8-37F6444B68BC.html | 18:02 |
| xibalba_ | Installing the NetApp® NFS Plug-in 1.0.20 for VMware® VAAI | 18:02 |
| patdk-wk | but scsi vaai commands, I don't see how that can exist on nfs | 18:02 |
| patdk-wk | heh, plugin :) | 18:02 |
| xibalba_ | yeah | 18:02 |
| patdk-wk | not real vaai, but a vaai emulator | 18:02 |
| xibalba_ | correct | 18:03 |
| patdk-wk | nope, I dropped all my netapps | 18:03 |
| xibalba_ | oh? | 18:03 |
| patdk-wk | but I notice, vmware when it svmotions a vmdk, it will just move the data, if it's thin, it just moves what you expect | 18:03 |
| xibalba_ | quick backwards tangent : in this doc, https://help.ubuntu.com/community/KVM/Installation , i see references to 10.04/9.10, is it outdated? | 18:04 |
| patdk-wk | but on nfs, it reads it as if it's thick, then writes it as thin | 18:04 |
| xibalba_ | ^ yes i hate this ! | 18:04 |
| patdk-wk | so a 1tb empty vmdk on nfs, takes forever :( | 18:04 |
| xibalba_ | it causes SDRS to take much longer than it should | 18:04 |
| patdk-wk | and my last issue with nfs, why I discounted it in my first attempts to use it | 18:04 |
| patdk-wk | was no sioc | 18:05 |
| xibalba_ | ah we've got SIOC on our NS | 18:05 |
| patdk-wk | I keep having a single vm, pounding the nfs disks, and causing the others not to get reasonable io | 18:05 |
| xibalba_ | though i dont think the SIOC is properly working | 18:05 |
| xibalba_ | as i've had the same observation as you | 18:05 |
| patdk-wk | well, I don't even know of esxi does that now, I did these tests in 4.0 was released :) | 18:05 |
| patdk-wk | iscsi is more to deal with, but it seems to solve my isuses, so I'm happy :) | 18:06 |
| patdk-wk | or I just use fc directly | 18:06 |
| xibalba_ | yeah the config is more involved, but its only one time | 18:06 |
| patdk-wk | two clusters use fc, one uses iscsi, and one uses das (with iscsi for just when upgrading) | 18:06 |
| patdk-wk | actually, since the multipath vmotion, it's really the same | 18:07 |
| patdk-wk | make 1 nic per port for vmotion and iscsi, repeat | 18:07 |
| xibalba_ | yup | 18:07 |
| xibalba_ | should KVM be mentioned here in the list: https://help.ubuntu.com/14.04/serverguide/virtualization.html | 18:08 |
| patdk-wk | it is | 18:09 |
| patdk-wk | linking to kvm would be useless | 18:09 |
| patdk-wk | as it's just a processer instructions | 18:09 |
| patdk-wk | pointing to qemu would be more useful, as it's the program that actually uses kvm | 18:09 |
| patdk-wk | but they link instead to the usermanagement interfaces to those tools | 18:10 |
| patdk-wk | being libvirt, or ubuntu cloud, ... | 18:10 |
| === mjohnson151 is now known as mjohnson15_2 | ||
| === adam_g` is now known as adaM_g | ||
| === adaM_g is now known as adam_g | ||
| zul | hallyn: your cgmanager patch fails with 1.2.8 | 19:17 |
| === etzsch is now known as markour | ||
| hallyn | zul: leave the patch commented out and sling over the .dsc | 19:49 |
| xibalba_ | im reading through this KVM doc, http://www.howtogeek.com/117635/how-to-install-kvm-and-create-virtual-machines-on-ubuntu/, and i'm looking for something thats more command line oriented. so i can run it on a server at my colo w/out having to do X11 fwding to my desktop | 20:39 |
| TJ- | xibalba_: virsh and libvirt might be what you're looking for | 20:41 |
| xibalba_ | k i found this and it looks more cli; | 20:42 |
| xibalba_ | http://www.howtoforge.com/virtualization-with-kvm-on-ubuntu-12.10 | 20:42 |
| xibalba_ | can i get more VMs out of my host w/KVM vs VirtualBox? | 20:43 |
| xibalba_ | since KVM has a shared kernel space | 20:43 |
| Seranok | how do i determine why a cURL request is taking 5-15 seconds from my server? | 21:21 |
| Seranok | it only takes < .2 seconds from my browser | 21:21 |
| Seranok | is there a request queue? | 21:22 |
| === Corey is now known as QuinnyPig | ||
| sarnold | Seranok: are your browser and curl running on the same machine? | 21:35 |
| === _thumper_ is now known as thumper | ||
| Seranok | sarnold: no | 23:26 |
| Hen0k | .. | 23:27 |
| sarnold | Seranok: does one have an IP that resolves quickly and the other have an IP that doesn't resolve quickly or doesn't resolve at all? | 23:27 |
| Seranok | sarnold: no, i tried pinging from both machines, takes the same amount of time | 23:28 |
| Seranok | that's why i'm wondering if there's some http request queue or i'm being throttled | 23:28 |
| sarnold | Seranok: I'm curious if your web server is trying to do reverse lookups on the IPs; from the server, run 'host <ip>' for each IP of your clients | 23:28 |
| Seranok | what would that show? | 23:30 |
| sarnold | Seranok: if one resolves instantly and the other doesn't resolve at all, then I'll say "go turn off domain name logging in your access.log" :) | 23:31 |
| Guest79575 | I have trouble with firewall I setup using UFW on 14.04 server on AWS/EC2: 'ufw status' firewall is enabled, rules look good to I can SSH into the machine, but on reboot I cannot get back in. So I stop the instance, detach the root volume, and re-attach and mount it on another machine where I disable UFW on startup, then re-attach to the first machine and boot it and SSH in fine, activate UFW and still SSH in fine, reboot and | 23:31 |
| Guest79575 | again I cannot SSH in. Did something change here between 12 and 14? | 23:31 |
| sarnold | Seranok: if both resolve instantly or both fail to resolve, it'll mean someone else needs to take a crack at your question :) hehe | 23:31 |
| Seranok | sarnold: i think that's the issue | 23:43 |
| === histo_ is now known as histo | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!