[00:14] I am looking to put a machine somewhere on the web (fast and reliable network connection) so that i can control and host webpages (and applications written in python). I am in the US and would not like to pay too much for just electricity and network connectivity. Has anyone faced this problem here? Any pointers to solutions? [00:15] pythonirc101: which colocation facilities have you looked at? [00:15] pythonirc101: have you looked at VPS? [00:16] i did not know about colocation , will look at VPS now [00:16] colo is expensive [00:16] leasing a VPS or even a dedicated server is usually much cheaper [00:16] yeah, i'd only go colo if you need to have your own hardware, which is seldom the case. [00:16] ^ [00:17] VPS or dedicated servers are ALWAYS better because they're cheaper [00:17] i'm speaking from experience there :P [00:17] what is VPS? [00:17] virtual private server. [00:17] virtual private server [00:17] basically... [00:17] expect a colo to cost 5-10 times more than a dedicated server [00:17] "the cloud"!!!!!! OMG [00:17] its a server that is run virtually off of a host node [00:17] haha [00:17] how much cost am i looking at? for electricity and network cable? [00:18] and a dedicated server to cost 10-30 times more than a VPS [00:18] pythonirc101: for colo? [00:18] i dont mind getting the hardware from the provider or paying for it myself. [00:18] at work, I'm in the process of decommissioning a colo box we pay $500/mo for [00:18] electricity, networking, power, rent of space in the facility, bandwidth, etc. [00:18] not to mention the actual hardware can cost horrid amounts [00:18] i bought a Dell PowerEdge 2500, cost me $2500 [00:18] then put that thing into colo [00:18] you can find dedicated servers starting around $100/mo [00:19] ^ [00:19] and VPSes, depending on how reliable you need them, anywhere from $5 up [00:19] ^ [00:19] decent VPSes i've seen are around, what $20 - $30 a month? [00:19] for reliable [00:19] yeah [00:19] TheEvilPhoenix: how much do you pay per month to put that dell in colocation? [00:20] pythonirc101: you mean how much DID I pay [00:21] linux VPSes also tend to cost a lot less than windows ones [00:21] it shifted between locations over 2 years [00:21] but first place was $450/month [00:21] some places offer the same pricing for both, but that's before you factor in 20 gigs of hard drive space wasted on windows itself [00:21] second was $49.99/month with horrid uptime reliability [00:21] finally said "Screw this", and cancelled the colo last week [00:21] TheEvilPhoenix: what did you do after that? [00:22] the server is now en route to my nice, static ip, business class internet hookup :P [00:22] yeah [00:22] pythonirc101: i'm waiting on the server to arrive, but its going onto my nice business class internet [00:22] TheEvilPhoenix: that costs me $150 for 10mbit line/month [00:22] in some cases it's actually cheaper to just buy a cable line and find someone who will let you store it there [00:22] pythonirc101: well i have an advantage... [00:22] pythonirc101: i dont run the business class internet [00:22] er [00:23] enterprise class, even [00:23] TheEvilPhoenix: Thats cool [00:23] one of the companies i work with does [00:23] and they are allowing me to set the server up there [00:23] free of charge [00:23] (partly because they ALREADY pay me flat-fees monthly for being on-call for when their servers explode) [00:24] yeah [00:24] the actual cheapest way to run a server is to know people [00:24] mhm [00:24] but the cheapest way that you can measure in dollars is VPS [00:25] mhm [00:25] hmmm [00:25] I need to find someone i know who has a good internet connection and is willing to let me colocate [00:26] how does one buy fiber connections? They probably are very expensive? [00:26] I know that there is a dead server right outside my building [00:26] another reason the cost is so high for colocation [00:27] if you're not running the latest greatest gear, you're also wasting a lot of electricity [00:27] qman__: i dont care about electricity for now...only one machine [00:27] it's just economy of scale on so many levels [00:27] I just want to buy a good machine and put it in a good place [00:27] but why, exactly, do you want to do that [00:28] or buy a good ethernet connection to my building...i somehow think that will be harder to do than to colocate [00:28] why is a VPS not good enough, and more accurately, why are 20 VPSes not as good? [00:28] because that's about the cost difference [00:28] qman__: how much would a vps cost me? [00:28] will they give me the latest and greatest machine with SSDs on them for the right price? [00:28] bottom end from reputable companies is $5/mo [00:28] they go up from there [00:28] I run web apps, not just web servers [00:29] then you probably want a dedicated server [00:29] you can get quad core machines with plenty of RAM and disk space for $100/mo [00:29] qman__: is there a vps provider that will let me be root on ubuntu and give me a good machine? [00:29] all of them, pretty much [00:30] any pointers? [00:30] don't look below the $100 mark for a dedicated server [00:30] http://www.vpsguides.com/vps-hosting-comparison/ [00:30] there are some cheaper than that but they're best avoided [00:30] for a VPS, if you care about performance, aim $20/mo and up [00:31] 30GB disk space on most, not an option...at leave 4TB or more disk space is what i need...and that should cost $160? [00:31] if you need that much space you'll have to pay more [00:31] but it's not unattainable in the dedicated server space [00:32] 2x2TB drives = $200 at most today [00:32] 30GB disk space is a joke... [00:32] you're not looking at all the angles [00:32] to keep the price down users share hardware, that's how a VPS works [00:33] it makes the most of the resources, and to do that, they have to limit disk space to allow for as many VPSes as the server can reasonably handle at once [00:33] that's why more space costs more money [00:33] i see [00:33] a VPS isn't the best choice if you have such a large data requirement, look at dedicated servers instead [00:34] qman__: or perhaps colocation [00:34] it's not likely you'll be able to colocate cheaper than a dedicated server, even with your requirements [00:34] unless you don't need datacenter quality accomodations and can stick it in someone's house with a comcast line [00:36] and just to give you a sense of the market [00:36] 4TB is _huge_ for a web server [00:37] hmmm [00:37] nobody's going to have a 4TB 'plan', you'll have to shop with extra disk addons [00:37] i do need a colocation [00:38] fair enough, but it's going to be expensive, no two ways around it [00:38] hmmm [00:38] I assume that by having 4TB of data, you actually want to serve that data at a pretty high rate [00:38] and the $500/mo rate I mentioned earlier is for 1Mbit/s averaged [00:38] unless i open up a colocation center myself and share the network bandwidth...the only expensive part is network... [00:39] And i know there is dead fiber in front of my building [00:39] its just a matter of if i can lease it... [00:39] dead fiber is usually dead for political reasons [00:39] not likely you'll be able to use it [00:39] you are probably right [00:39] the company owns it and wants to sit on it... [00:40] no clue why [00:40] is verizon fios reliable enough to run a server on it? [00:41] that depends entirely on what kind of reliability you demand [00:41] it's roughly the same as business class cable [00:41] it's not measured in nines [00:42] hmmm [00:43] verizon is promising $100 for 25Mbps/up/down [00:43] http://smallbusiness.verizon.com/products/internet/fios/overview.aspx#plans [00:47] and we all know verizon fails :P [01:31] anyone seeing postgrey segault? jaunty [01:32] I'm using postgrey on lucid without problems [01:32] strace it? [01:32] sorry, I was mistaken, I'm on lucid as well. [01:34] twb`: I can't make any sense of this. Can you? -> http://pastebin.com/ZmGtua2t [01:35] bc: nothing obvious there [01:42] twb`: I purged and reinstalled and it works. Apaprently something in the configuration file was being incorrectly handled. [01:42] And apparently I can't work my hands. Keyboard. How does it work? [01:44] RoAkSoAx, your changes to orchestra look great. [01:48] bc: sorry, man, no idea [01:48] bc: if you have etckeeper installed, you can see what the differences were between working and broken config [01:56] smoser thanks [01:57] smoser.waiting fpr the branch to get merged into juju so i can release [01:58] hey guys, i need some advice on a server setup. ive got an old P4 with HT and 3 gb of ram im thinking of putting server 10.04 lts on it and serving up a drupal site as well as hosting a minecraft server. i only have a 40 gb hdd and im trying really hard to make it as redundant as possible I was first thinking about seting up virtualbox and just doing periodic back ups of the vdi but i didnt want to waste more resources. s [01:58] o i am not thinking of just installing ubuntu straight up. does anyone have any good ideas on how to keep a catastrophy to a minimum and how to quickly recover once this old dell finally eats it? [02:02] T3CHKOMMIE: not running minecraft alongside anything else is a good start [02:03] T3CHKOMMIE: minecraft servers require MINIMUM 2GB to run. [02:03] to run effectively* [02:04] TheEvilPhoenix, i have a XBMC install running all this crap right now and the mini ITX board can only support 2 gb... so im going to have to move to the P4 i was just wondering if anyone new of a way for me to keep this thing backed up up-to-the-day so that if... i mean... WHEN it crashes my site and other services will only be down for a short time... any tricks? [02:04] ideally i wanted to be running all this stuff on a NICE esxi box... but alas, i dont have the $$$ for that kinda setup just yet. === medberry is now known as med_out === twister004_ is now known as twister004 [03:25] twb`: no problem, all good. thanks for help [03:25] np === twister004_ is now known as twister004 === almaisan-away is now known as al-maisan === al-maisan is now known as almaisan-away [05:53] hi all === CrazyGir is now known as Guest16661 [06:37] New bug: #851611 in eucalyptus (universe) "Oneiric: Upgrade to 2.0.1+bzr1256 blocks" [Undecided,New] https://launchpad.net/bugs/851611 [06:56] New bug: #851625 in munin (main) "snmp__if plugin does not work" [Undecided,New] https://launchpad.net/bugs/851625 === twister004_ is now known as twister004 === twister004_ is now known as twister004 [07:53] morning all === sanderj is now known as Sander^work [09:31] Daviey: could you have a look at https://code.launchpad.net/~tpatil/nova/lp782364/+merge/75259 [09:31] Daviey: if you ok with it, I'll merge it in the Essex packaging branch [09:31] Daviey: that said, we consider adding that to diablo too, since the option seems useful if you actually wants FloatingIPs to be usable [09:32] iputils-arping is in main so damage is limited, but wanted your opinion first [09:43] Daviey, SpamapS: do we now have a working eucalyptus in universe? [09:48] jamespage: seems so, testing would help.. :) [09:48] ttx: looking [09:49] Daviey: right - I need to test it for a proposed upgrade to libasm3-java to support a late FFE for eclipse 3.7 [09:49] Daviey: I've never actually installed it from anywhere other than the ISO - where do I find instructions [09:49] (and good morning BTW) [09:49] morning. [09:50] jamespage: https://help.ubuntu.com/community/UEC/PackageInstall [09:50] ta [09:51] fmd [09:54] ttx: landing in the Essex branch is no problem.. if it's going to be backported to Diablo, then it looks OK.. and not even enabled by default. [09:54] yep [09:55] ok, I'll make it happen [09:59] make it so. [10:00] http://bit.ly/nzpN73 [10:05] * jamespage restarts his install [10:06] network installs don't like it when the power goes off [10:46] is it possible that upgrading from 10.10 server "automagically" installed XEN and spontaneously configured GRUB to boot from a DOM0 that I haven't obviously configured (not being aware of it)? === nandemona is now known as nandemonai [10:53] Daviey: hmm it was reticent to startup without a kick [10:54] jamespage: what sort of kick? [10:55] Daviey: its hanging during startup - "wget -q -T10 -t1 -O- --no-check-certificate https://192.168.1.52:8443/register" [10:55] during initial configuration [10:56] hmm [10:56] I think its when the various components try to "restart eucalyptus" in postinst [10:56] this happend once before.. [10:59] I can't get a response from https://192.168.1.52:8443/register either [11:03] Daviey: any pointers on that one [11:05] jamespage: Honestly, i cannot remember.. I think one of the components didn't quite start properly. [11:05] I'm sorry i'm not being more help. [11:05] Sounds a bit like bug 503180 [11:05] Launchpad bug 503180 in eucalyptus "[SRU] eucalyptus-cloud doesn't reply to requests (eucalyptus doesn't work after reboot or services restart issues due to upstart networking behavior)" [High,Fix released] https://launchpad.net/bugs/503180 [11:06] Hmm [11:06] but I don't think its the same issue [11:06] not sure it is, but could well be [11:12] Daviey: it gets there in the end by waits for 60 x 10 seconds before giving up [11:14] jamespage: Hmm.. maybe best to defer it and speak with upstream later on today. [11:40] Daviey: hmm - odd that I can get that URL using chromium from my laptop [11:40] but not using wget or curl [11:43] jamespage: wait, it doesn't working by using curl on your laptop? [11:43] or just not from the dev box, accessing localhost? [11:44] either - I get the same thing from curl on both the euca install - localhost - or remotely [11:44] works OK in firefox as well [11:44] * SSLv3, TLS handshake, Client hello (1): [11:44] * Unknown SSL protocol error in connection to 192.168.1.52:8443 [11:44] * Closing connection #0 [11:44] curl: (35) Unknown SSL protocol error in connection to 192.168.1.52:8443 [11:46] hmm - I don't think this is actually euca [11:46] Daviey: ^^ - if I run the wget command from a natty chroot - it works just fine [11:47] but not from oneiric [11:48] jamespage: ahhhhh [11:48] We dropped SSLv2 support this cycle... i wonder if it is that? [11:48] 12:44 < jamespage> * SSLv3, <-- bah [11:51] New bug: #829234 in socat (universe) "[MIR] socat" [Undecided,In progress] https://launchpad.net/bugs/829234 [11:53] Its using SSLv3 [11:55] * jamespage scratches his head [11:58] Daviey: ignore my natty observation - think that was red herring [11:58] but it did work - think that euca just unfroze or something === Ursinha-afk is now known as Ursinha [12:12] i setup my login. i see it. but i cannot be accepted? [12:12] http://i.imgur.com/J9DdG.png [12:13] i also did ubuntu-prosper [12:14] wtf. where is the ubuntu serverh andbook [12:17] !serverguide | Goatesper [12:17] Goatesper: The Ubuntu server guide may be found at http://help.ubuntu.com/10.04/serverguide/C/ [12:17] i see. i did admin-prosper [12:17] that was the minor difference [12:17] 10.04 is same as 11.04? [12:18] -bookmark [12:18] Goatesper: No, but 10.04 is... [12:18] meh. [12:35] SpamapS, give me a ping when you start today please [12:36] trying to verify that euca might just be usable in oneiric [12:37] jamespage: SpamapS is the new expert for euca btw :) [12:37] really? how interesting [12:37] I saw he applied the upstream patch yesterday [12:38] jamespage: Yeah, that made it his problem :) [12:39] hehe [12:49] morning [12:49] sucks to be him :) [12:49] afternoon zul [12:50] Daviey: sounds like a blame. [12:50] Daviey: no its morning [12:54] What's the preferred API entry point to have Python API access to openstack? euca2ools.commands? boto? [12:54] depends on what you want to do [12:55] I want API access to stuff like euca-run-instances [12:55] boto [12:55] OK cool, thanks [12:55] +1 on boto [12:55] * rbasak is familiar with boto already, so that's convenient :) [13:04] rbasak: if you need concurrency, use txAWS. [13:04] .. which i am still trying to get my head around [13:07] ive seen problems with txaws against openstack though [13:08] zul: oh? Where? [13:08] with ensem...i mean juju [13:09] zul: I think SpamapS fixed that yesterday [13:09] probably [13:37] hi [13:38] does anybody know about posibilities to enable hugepages in EC2 cloud? [13:57] New bug: #851900 in eucalyptus (universe) "Eucalyptus slow to startup with broken connections to :8443/register" [Undecided,New] https://launchpad.net/bugs/851900 [14:05] Daviey, around ? [14:06] smoser: no. [14:06] err, yes [14:07] so... [14:07] nova [14:07] what is the plan for nova and beta ? [14:07] are we expecting an upload prior to? [14:07] smoser: as in a new snapshot? [14:07] yeah [14:07] yah [14:07] k [14:07] im uploading a version as we speak [14:07] zul i believe is driving that [14:07] ah. [14:07] well. [14:07] i'd like to ask for one on tuesday [14:07] :) [14:08] why tuesday? [14:08] cause thats later than today [14:08] nova releases on thursday [14:08] we release on thursday, so thats too late [14:08] :) [14:08] ok i can do one on tuesday as well [14:09] http://paste.ubuntu.com/690824/ [14:09] zul: can you include bug numbers for bugs with a ubuntu task? [14:09] thats the list of stuff that i'm racing to get in. [14:10] smoser: you are a rock star [14:10] if someone wants to write some python code to say "is this qcow image compressed" , that would rock [14:10] smoser: jeebus [14:10] anyone ? [14:10] bueler ? [14:11] smoser: heh ok...i have at least one fix to get in as well [14:11] smoser: can you produce two small qcow files? [14:13] you mean for you to test with ? Daviey ? [14:13] yeah, its easy [14:13] smoser: yah [14:15] truncate --size 100M my-raw.img [14:15] mkfs.ext3 -L himom my-raw.img -F [14:15] qemu-img convert my-raw.img -O qcow2 my-qcow2.img [14:15] qemu-img convert my-raw.img -O qcow2 -c my-qcow2-compressed.img [14:16] $ ls -l my-*.img [14:16] -rw-r--r-- 1 smoser smoser 344576 2011-09-16 10:15 my-qcow2-compressed.img [14:16] -rw-r--r-- 1 smoser smoser 3014656 2011-09-16 10:15 my-qcow2.img [14:16] -rw-rw-r-- 1 smoser smoser 104857600 2011-09-16 10:15 my-raw.img [14:16] i can give you some logs of conversation i had in #kvm also regarding it. [14:16] smoser: fancy dropping it all in a mail? [14:16] easier for me to put it together. :) [14:17] sure. [14:17] ta === almaisan-away is now known as al-maisan [14:21] Daviey, http://paste.ubuntu.com/690838/ [14:22] if you want to mail that to yourself, go ahead. [15:00] Daviey smoser adam_g zul kirkland FYI, I'm changing everything that says ensemble to juju in Orchestra that will be released later today [15:01] ack [15:01] so if something breaks just let me know ;) [15:01] New bug: #851963 in vsftpd (main) "vsftpd with likewise-open doesnt create directory on login" [Undecided,New] https://launchpad.net/bugs/851963 [15:02] RoAkSoAx: good luck [15:03] RoAkSoAx: rocking === med_out is now known as medberry [15:26] jamespage / zul: bug 825670 fix has landed in oneiric now? [15:26] Launchpad bug 825670 in swift "python-swift: remove the user on purge" [Medium,Fix committed] https://launchpad.net/bugs/825670 [15:27] Daviey: it should have yes [15:27] zul: Are you working on bug 802402? [15:27] Launchpad bug 802402 in image-store-proxy "convert to dh_python2" [Undecided,Confirmed] https://launchpad.net/bugs/802402 [15:27] Daviey: no i havent i can figure that out now [15:29] * Daviey eats [15:30] * RoAkSoAx hungry [15:34] It is possible to classify vlan traffic with u32 ? [15:36] New bug: #825670 in swift (main) "python-swift: remove the user on purge" [Medium,Fix committed] https://launchpad.net/bugs/825670 [15:46] smoser: SpamapS: so should libvirt-bin switch from starting on 'stopped networking RESULT=ok' to using static-network-up? [15:47] I'm thinking it needs to. And I'm not sure what we do about SRU for that (since static-network-up doesn't exist in natty and earlier) [15:51] genii-around: see http://brownian.org.ua/?p=189&langswitch_lang=en if it applies to your problem [15:52] just-a-visitor: Reading, thanks [15:53] no, wait [15:53] I think from oneiric onward we can drop that from libvirt [15:54] bc it starts on runlevel, but rc-sysinit starts on static-network-up, so libvirt will now automaticlaly wait for all netdevs to be up? [15:54] SpamapS: jhunt: ^ [16:03] Daviey: whos the best person to talk to upstream from euca? [16:10] I've a machine at home and would like to run apache from here (dynamic ip). Does anyone know to point my dns to this apache box? I guess i'll have to do port forwarding...I use zoneedit for dns for now. [16:14] I guess I'll try to use ddclient [16:15] hallyn: +1 thats what I would like to see from most daemons.. 'start on runlevel [2345]' [16:16] SpamapS: ok, i opened bug 852000 for that. [16:16] Launchpad bug 852000 in libvirt "libvirt no longer needs to start on stopped networking" [Undecided,New] https://launchpad.net/bugs/852000 [16:16] SpamapS: but we missed freeze. not sure we can sneak something like this in for o at all any more? [16:17] New bug: #852000 in libvirt (main) "libvirt no longer needs to start on stopped networking" [Undecided,New] https://launchpad.net/bugs/852000 [16:18] SpamapS: so is there any way that static-network-up will get SRUd? [16:19] hallyn: the event, maybe, the rc-sysinit change, never. [16:19] the event would be ok, then libvirt could start on that [16:20] bc start on stopped networking is insufficient [16:22] SpamapS: was all the work for static-network-up contained within the ifupdown package? === al-maisan is now known as almaisan-away [16:24] jamespage: grazino [16:27] hallyn: no, upstart has the failsafe and rc-sysinit changes [16:27] hallyn: the event though, yes, is self contained in ifupdown [16:28] jamespage: you rang? [16:28] SpamapS: hey - yes [16:28] although I may have answered by question since [16:29] I wanted to check what testing you did of the eucalyptus upload yesterday [16:29] I've been testing today and its distinctly unhappy [16:29] SpamapS: is the failsafe change to work around the rc-sysinit change? I.e. if we don't have rc-sysinit change but we have static-net-up we don't need the failsafe, right? [16:30] its not your upload BTW [16:34] SpamapS: ok, full disclosure :) are you already planning on doing the static-net-up fix for lucid 10.04.4 for bug 580319 ? [16:34] Launchpad bug 580319 in upstart "init.d controlled services launch before all interfaces are up, thus failing to start" [Undecided,Confirmed] https://launchpad.net/bugs/580319 [16:42] All right I think I have to take a stab at it === cloakable_ is now known as cloakable [16:48] hallyn: I don't think we're going to be able to SRU that. Its a massive change in behavior. [16:49] jamespage: I smoke tested euca.. started it up, was able to login to the clc .. did a couple of reboots.. not much else [16:49] SpamapS, hrm - interesting [16:50] SpamapS: I'm hitting this issue - and so is TeTeT during upgrades [16:50] wget can't talk to the clc over http://xxx:8443 [16:50] chromium and firefox can [16:51] bug 851900 [16:51] Launchpad bug 851900 in eucalyptus "Eucalyptus slow to startup with broken connections to :8443/register" [Undecided,New] https://launchpad.net/bugs/851900 [16:52] jamespage: this is wget on both localhost and your laptop? [16:53] Daviey: yep [16:53] sometimes you get no reponse, sometimes a partial and sometime full [16:53] but most times no response [16:54] jamespage: i wonder if this is a bridge issue? [16:54] maybe X based browsers are more aggresive in trying? [16:54] Daviey: same issue on localhost [16:55] I did sniff the network packets - they looked OK === lullabud is now known as warzauwynn [16:55] jamespage: Yeah, but does that actually use lo directly? [16:55] i.e. no unexpected RST's etc... [16:55] hmm. [16:55] Daviey: I think so [17:02] how do you control which TTY is selected when the system boots up? one of my machines is always going to a blank tty, no clue why. [17:03] warzauwynn: vt.handoff=# [17:03] warzauwynn: in the kernel args in grub [17:03] w00t, thanks [17:04] SpamapS: Is juju naming transition complete? [17:04] Daviey: trunk has been renamed [17:04] Daviey: I'm preparing a FFe for the renaming of the binary packages and such, to also be timed with landing a couple other things we already planned to land late. [17:06] hallyn, surely we're allowed to fix bugs still [17:06] (regarding "SpamapS: but we missed freeze. not sure we can sneak something like this in for o at all any more?") [17:07] Daviey: I subscribed graziano to the bug report and requested some help [17:10] smoser: which one is that in referecne to? [17:11] smoser: the oneiric one is just a cleanup - no misbehavior without it [17:11] smoser: the other stuff is only for natty and earlier [17:11] oh. i thought you were saying for oneiric. [17:11] we've long missed "freeze" for natty [17:11] i think that was about 6 months ago [17:14] * RoAkSoAx will be back in an hour [17:15] smoser: i must have missed that email [17:15] j/k [17:15] SpamapS: weird, changing the TTY didn't help… whenever i boot up, regardless of which TTY is selected it always starts with a blank window. I have to switch away and back to get any text to display on that tty. any ideas? [17:16] smoser: do you think that this part; [17:16] * invoke dhclient3 with '-1', meaning it should exit failure if it does [17:16] not receive a response in 60 seconds. (LP: #838968) [17:16] warzauwynn: no, that is weird. What release? [17:16] should be part of an SRU for static-network-up? [17:17] it's 11.04, been upgraded through at least one version of 10, maybe starting at 9. [17:17] it's been a problem for a while and i've always just dealt with it… [17:17] not like it's a huge problem either, just kinda odd. [17:18] hallyn, i would think you should not do that. [17:18]