Back to Home / #linode / 2006 / 05 / Prev Day | Next Day
#linode IRC Logs for 2006-05-09

---Logopened Tue May 09 00:00:03 2006
00:00|-|afv-13 [] has quit [Ping timeout: 480 seconds]
00:16|-|fake [] has joined #linode
00:28|-|spr [] has quit [Quit: Spoon!]
01:05|-|Newsome [] has quit [Quit: Linux: Now with employee pricing!]
01:19|-|MoonRanger [] has quit [Ping timeout: 480 seconds]
01:51|-|nybble [] has quit [Quit: Leaving]
02:32|-|md|cm [] has joined #linode
02:49|-|dryan [] has quit [Quit: Leaving]
02:56|-|Battousai [] has quit [Ping timeout: 480 seconds]
04:28|-|womble [] has quit [Ping timeout: 480 seconds]
04:34|-|k4nzer [~k4nzer@] has joined #linode
04:35|-|k4nzer [~k4nzer@] has quit [Quit: ]
04:53|-|md|cm [] has quit [Quit: Got Itch?]
05:16|-|Eman [] has quit [Read error: Connection reset by peer]
05:36<Dreamr3>i just upgraded my memory on my linode...
05:37<Dreamr3>anyone here?
05:39[~]encode is here
05:41<Dreamr3>what folder do i need to move with 2.6?
05:42<encode>you mean the /lib/tls folder?
05:42<Dreamr3>yeah that
05:44<Dreamr3>ha, failed staring mysql?
05:44<Dreamr3>must be my lucky day
05:45<Dreamr3>ok, lets try this again
05:47<Dreamr3>mikegrb: ping
05:47<Dreamr3>my interfaces won't come up under 2.6 :(
05:47<encode>what kernel are you using?
05:48<encode>theres a few different 2.6 kernels
05:48<encode>from memory the 2.7.17? has some issues
05:49<encode>maybe try one of the other 2.6 kernels
05:49<encode>other than that i dont really know, sorry dude
05:49<Dreamr3>you use 2.6?
05:49<encode>i probably will next time i reboot
05:50<Dreamr3>guess it's back to 2.4
05:50<Dreamr3>i really don't have time to screw around
05:55|-|flatronf700B [~flatronf7@] has joined #linode
06:58<Dreamr3>caker: you're not overselling RAM now, are you?
07:22|-|Eman [] has joined #linode
07:52|-|Antitribu [] has joined #linode
07:53<Antitribu>anyone awake have much experience with sftp?
08:02|-|Quizzer [] has joined #linode
08:56|-|jekil [] has joined #linode
09:29|-|Dreamrr_3 [] has joined #linode
09:34|-|Dreamr3 [] has quit [Ping timeout: 480 seconds]
10:35<linbot>New news from forums: Apache2 segfaults (ubuntu 5.10, latest 2.6 series) in Feature Request/Bug Report <>
10:36|-|spr [] has joined #linode
11:01|-|FireSlash [] has joined #linode
11:02|-|vodka [] has joined #linode
11:11|-|jekil [] has quit [Ping timeout: 480 seconds]
11:20|-|wdestefano [] has joined #linode
11:41|-|FireSlash [] has quit [Quit: Leaving]
11:59<linbot>New news from forums: Distro: Ubuntu 4.10 now available in Announcements <>
12:05|-|Marcel [] has joined #linode
12:30|-|afv-13 [] has joined #linode
12:38|-|odedra [] has joined #linode
12:43|-|Battousai [] has joined #linode
12:43|-|Battousai [] has quit [Quit: ]
12:54|-|trinchante [~trinchant@] has joined #linode
12:54|-|trinchante [~trinchant@] has quit [Remote host closed the connection]
13:10|-|FireSlash [] has joined #linode
13:15|-|jekil [] has joined #linode
13:41|-|FireSlash [] has quit [Quit: Leaving]
14:16<linbot>New news from forums: Apache2 segfaults (ubuntu 5.10, latest 2.6 series) in Linux, Apache, Mysql and PHP (LAMP) Forum <>
14:18|-|Spads [] has quit [Remote host closed the connection]
14:22|-|Spads [] has joined #linode
14:36<@caker>Dreamrr_3: no
14:38<afv-13>what was the question?
14:39<@caker>do we oversell ram
14:44<Quizzer>quick and probably very beginner (newbie) question ..
14:44<Quizzer>in Centos; how do you create another user so i dont have to always use root
14:46<afv-13>useradd -m <username>
14:46<afv-13>and then passwd <username>
14:46<Quizzer>oh cool thanks ..
14:47<afv-13>and then disalow root logins with ssh
14:47<Quizzer>very much a newbie and new customer here .. but its an awesome service
14:48<afv-13>it's a great place to learn
14:49<afv-13>if you screw up you can have a fully working install within minutes
14:50<Quizzer>yep ..
15:11<gpd>Quizzer: adduser username is easier to use than useradd
15:11<gpd>it sets up the home dir and a bunch of other stuff for you
15:12<Quizzer>ok ..
15:14|-|Redgore [~Redgore@] has joined #linode
15:14|-|FireSlash [] has joined #linode
15:17|-|Redgore [~Redgore@] has quit [Quit: ]
15:32<Quizzer>gpd: actually dont have the adduser command .. but the useradd worked
15:35<gpd>caker - in LA... death in minutes I expect ;)
15:36<gpd>Quizzer: sux to be you - should have gone with Debian|Ubuntu (/me runs)
15:36<Quizzer>probably should have
15:36<Quizzer>but I already have a ubuntu box .. and thats running on my DSL connection at home on an old laptop
15:37<Quizzer>i wanted something different
15:37<gpd>err -- if you say so!
15:38<gpd>i have complete ubuntification (desktop + linode) - my assimilation is complete
15:39<gpd>even the new $10 have gone brown -- clearly you will not be far behind
15:40<gpd>resistance is useless (!)
15:40[~]gpd takes more medication
15:55|-|thoth39 [] has joined #linode
16:03|-|FireSlash [] has quit [Ping timeout: 480 seconds]
16:20|-|wdestefano [] has quit [Quit: CGI:IRC (Session timeout)]
16:50|-|MarcelH [] has joined #linode
16:56|-|Marcel [] has quit [Ping timeout: 480 seconds]
17:03<Dreamrr_3>ubuntu is alright
17:11<fo0bar>gpd: OMFG! Are you running the new Incandescent Iguana build?!?
17:13<gpd>zOMFG! almost! I'm like totally excited about it - SUPER!
17:18|-|spr [] has quit [Quit: "Ah, the bounce has gone from his bungie"]
17:27|-|Lego [] has joined #linode
17:27|-|Antitribu [] has quit [Read error: Connection reset by peer]
17:31|-|Antitribu [] has joined #linode
17:31|-|Lego [] has quit [Read error: Connection reset by peer]
17:38|-|MarcelH [] has quit [Quit: Leaving]
17:39|-|Antitribu [] has quit [Ping timeout: 480 seconds]
17:43|-|Antitribu [] has joined #linode
17:51|-|besonen [] has joined #linode
17:52|-|besonen_ [] has quit [Read error: Connection reset by peer]
17:56|-|gpd [~gpd@] has quit [Quit: Reconnecting]
17:57|-|gpd [~gpd@] has joined #linode
18:00|-|jekil [] has quit [Ping timeout: 480 seconds]
18:06|-|FireSlash [] has joined #linode
18:31<@caker>there aren't TTLs for individual records in bind/named, right? Just for the entire zone
18:31<@caker>oh, there are .. cool
18:32<@caker> $TTL IN A
18:33<@caker> $TTL IN MX 10
18:33<efudd>yes, you can have
18:33<thoth39>I think I may throw the towel fighting this xennode.
18:33<thoth39>I admit I'm pushing it, though.
18:34<@caker>thoth39: ?
18:34<gpd>efudd: back from SIN already?
18:34<gpd>How is SIN for you?
18:34<thoth39>It eventually stopped responding. No abnormal message in the log, but no response on lish or ssh.
18:35<efudd>waiting on my ride to the office
18:35<gpd>efudd: yes - it was like that wehn i was there... crazy hot.
18:35<thoth39>I'm trying to reproduce the system to learn something.
18:35<thoth39>What I did was basically upgrade the FC2 image all the way to FC4.
18:36<@caker>thoth39: your console shows it waiting at login
18:36<thoth39>Ah, yes, now it is, I'm redoing things to see where it breaks.
18:37<thoth39>It started with openldap failing for some reason; no message in the logs. That's all I found out.
18:38<tierra>caker: has there been any reports of instability with the 2.6.15-linode16 kernel? My node just dies with what looks like an alloc_pages failure, but I'll still have half of my swap free (linode 160 with 256 swap) when it happens
18:39<tierra>it tends to happen during high connection count / high traffic
18:41<tierra>I haven't tried stressing the other kernels yet, just thought I'd ask first
18:42<@caker>tierra: not that I'm aware of
18:46|-|taupehat [] has quit [Ping timeout: 480 seconds]
18:46|-|TheFirst [] has quit [Ping timeout: 480 seconds]
18:46|-|Quizzer [] has quit [Quit: Anonymous Ubuntu!! User...]
19:09|-|Quizzer [] has joined #linode
19:36|-|TheFirst [] has joined #linode
19:56|-|FireSlash [] has quit [Quit: Leaving]
20:07|-|kiomava [] has joined #linode
20:07<kiomava>host56 down
20:08<thoth39>I'm on it, and my node is up.
20:09<kiomava>can you log in to the console?
20:09<thoth39>Lish? There is something going on with lish, ssh is out, we're using telnet instead.
20:09<kiomava>i get connection refused trying to connect to
20:09<thoth39>Yeah, I know. :|
20:09<kiomava>ok i'm in, some strange message
20:10<thoth39>Reset your password afterwars, telnet is not secure.
20:10<kiomava>that's for sure
20:13<kiomava>hmm, networking stuff is down just for my node
20:13<kiomava>i can ssh localhost, but i can't ssh from a remote host
20:13<kiomava>can't ping
20:13<TheFirst>reboot fixes
20:13<kiomava>gunna try that
20:28<kiomava>ya that did it
20:28<kiomava>does caker know about this problem?
20:29<thoth39>I got this information via support ticket.
20:29<thoth39>The ticket is still open; I suppose it will close only when ssh comes back.
20:29<thoth39>Maybe you should open a ticket yourself, if just to monitor the status of this situation.
20:32<kiomava>just for the xen beta, not too worried
20:33<kiomava>long as he knows about it
20:33<kiomava>for the production linodes i'd be worried
20:34<thoth39>Yeah, yeah.
20:37|-|taupehat_ [] has joined #linode
20:37<taupehat_>10 ( 66.298 ms 199.085 ms 278.039 ms
20:37<taupehat_>where's host56?
20:39<taupehat_>where's host56?
20:40<linbot>New news from forums: Xen host down? in Xen Public Beta <>
20:40<thoth39>host56's ssh service is out.
20:41|-|taupehat [] has joined #linode
20:41<thoth39>host56's ssh service is out.
20:41<taupehat>what happened - HE portblock it?
20:41<thoth39>telnet is up for a while.
20:42<taupehat>or did the host just fall over, as would seem likely
20:42<thoth39>This has been going on for some days.
20:42<taupehat>not for me it hasn't
20:42<thoth39>Maybe sshd is interacting badly with the xen0 kernel. I don't know.
20:42<TheFirst>the ssh lish has been down for a week
20:42<TheFirst>when caker tries to restart it segs
20:43<taupehat>I've not been doing anything with lish - but everything died on mine in the past hour or so
20:43<taupehat>imap, ssh (direct to node, not lish), www, etc.
20:43<taupehat>I think the whole thing just went titsup
20:43<TheFirst>reboot fixes it
20:43[~]taupehat did a support ticket magic
20:43<TheFirst>statusd or something disconnected network or something (or so caker said)
20:44<taupehat>reboot needed =]
20:45<@caker>21:11 < jdike> caker, that's the network interrupt failing to allocate memory
20:45<@caker>21:11 < jdike> that's a particularly bad scenario
20:45<@caker>21:20 -!- jdike [] has quit [Quit: Leaving]
20:45<@caker>I'm guessing that it wasn't UML's fault
20:46<taupehat>so what's the good news? =]
20:46<fo0bar>caker: you need to get a higher SLA from jeff
20:53[~]taupehat makes frowny-faces and does not know...
21:04|-|taupehat1 [] has joined #linode
21:05|-|taupehat [] has quit [Quit: Lost terminal]
21:05|-|taupehat1 changed nick to taupehat
21:05<taupehat>rebooting often helps
21:07<taupehat>that was interesing
21:08<taupehat>I quit irssi while it was loading channels/servers and it coredumped
21:08<taupehat>me@taupehat:~$ irssi
21:08<taupehat>*** glibc detected *** free(): invalid pointer: 0x08112f28 ***
21:08<taupehat>Aborted (core dumped)
21:08<taupehat>d'oh =]
21:19|-|kiomava [] has quit [Quit: CGI:IRC (Ping timeout)]
21:19<thoth39>Everything fine with FC4 so far.
21:19<thoth39>I wonder if my problem wasn't really specific to openldap.
21:20<@caker>WRT the managed DNS thing, I suppose if we support slaving off user-run name server, when you go to add a zone you'll be able to choose if it's a master or a slave, and provide the slave details there...
21:21<@caker>also, I think it would be useful for people to be able to slave off our name servers, so they'll be a section where you can provide IPs of remote nameservers to allow to allow zone transfers to
21:21<fo0bar>caker: nifty
21:21<fo0bar>free service, or $29.95/zone/month?
21:22<@caker>besides the actual data entry forms, the interface look useable?
21:22<@caker>fo0bar: included with linode service
21:22<thoth39>Nice. Consider also support for TXT records; at least SPF uses them.
21:22<@caker>I guess nothing would prevent someone from using this for zones that have nothing to do with their Linode ....
21:22<@caker>thoth39: yeah, I had that in my notes .. I'll add it
21:23<gpd>caker: how many domains?
21:23<fo0bar>caker: looks good. but if you can spare it, I would recommend using different, authoritative-only DNS servers for service, and have the existing servers given out by DHCP as resolvers only
21:23<@caker>fo0bar: that's already the plan
21:24<fo0bar>also: SRV records
21:24<@caker>they won't be white-label, however .. it WILL say ""
21:24<thoth39>SRV records are a special type or record? I thought they were an application of TXT records.
21:24<@caker>fo0bar: hmm, ok
21:24<thoth39>..of record...
21:24<fo0bar>caker: I wasn't worried about that... when zone issues crop up and you're using a server both for authoritative and resolver, sometimes it makes diagnosing difficult
21:25<fo0bar>thoth39: nope
21:25<@caker>fo0bar: yeah, I was just adding that, because I know it'll eventually come up
21:25[~]caker forces linode down people's throats
21:25<thoth39>I see. I never got to experimenting with those records.
21:25<thoth39>When I read the RFC, it sounded great.
21:25<fo0bar>$ host -t srv
21:25<fo0bar> SRV 5 0 5222
21:25<gpd>caker: do you have defaults for the SAO times? or the range recommended?
21:25<fo0bar>that's one of the needed records for a proper jabber server
21:26<@caker>gpd: they'll be defaults, and checked to be within the normal ranges
21:26<@caker>I haven't decided if it'll be drop-down or an input field
21:26<gpd>so how many domains total? - I have 9 on my box - i know others ahve more
21:27<@caker>I dunno ... I haven't considered a limit at this point
21:27<gpd>nice talkin' to ya :)
21:27<@caker>probably no limit, or a very high number
21:28<fo0bar>maybe log queries, and set a monthly quota, like 1 million queries per month or something (I have no idea what an "average" would be)
21:28<@caker>fo0bar: yeah, some type of metering will be needed .. also what would be nice are notifications of zone transfers, etc
21:29<gpd>how about nameservers on separate class C's - did you mention that?
21:29<@caker>fo0bar: <-- accurate list of fields required?
21:29<@caker>gpd: that's a given, they'll be across the datacenters
21:29|-|Dreamr3 [] has joined #linode
21:30<fo0bar>caker: yes, but Service and Protocol are effectively collapsed (IE, _ass-scratcher._udp under the zone)
21:31<@caker>fo0bar: ok, so maybe in the UI an input for _ass-scratcher. and a drop down for udp or tcp, but save together in the record...
21:32<fo0bar>you could make it all user friendly (just ask for the service and protocols, add the leading underscores if they didn't provide them, etc), but most likely if a user is entering a SRV record, they know what needs to be done
21:32<fo0bar>caker: that could work
21:32<fo0bar>I've never seen anything other than _tcp and _udp
21:32<@caker>I have to decide how to handle when people enter just the host, periods, etc
21:33<@caker>I'd like to make the UI pretty smart about that stuff
21:33<@caker>like auto-expanding that stuff
21:33<fo0bar>god, I wish SRV took off. if browsers used SRV to look up web sites, it would effectively eliminate the need for most SLBs/GSLBs
21:33<@caker>If one word hostname, without a period, add the zonename; if a host without a period, look for other periods (IOW, they entered the entire thing in), etc
21:33<thoth39>Maybe you'll get good feedback on this design if there were a page in the Wiki about it.
21:33<fo0bar>ie or something
21:34<thoth39>fo0bar, true. www in is so 90's
21:34<fo0bar>thoth39: well, obviously you'd continue to type "" in the browser, the browser would take care of the SRV negotiation for you
21:35<@caker>maybe this would be a good time to do some AJAX stuff
21:35<TheFirst>you've already got managed dns set up?!
21:35|-|Dreamrr_3 [] has quit [Ping timeout: 480 seconds]
21:35<@caker>TheFirst: just a rough UI
21:36<TheFirst>ah...what ya gonna use...bind?
21:36<@caker>what else is there? :)
21:37<TheFirst> know the other ;)
21:37<fo0bar>ALL HAIL DJB!
21:37<TheFirst>but i'll refrain from comment seeing as mikegrb is in the room
21:37<fo0bar>does djbdns require patches even to compile on libcs from the last 5 years, following qmail's suit?
21:37<@mikegrb>mmm cake
21:38<@mikegrb>mmm cake
21:39<taupehat>but no rifk =[
21:39<TheFirst>all those idiot scripts need to be shoved where the sun don't shine
21:39<fo0bar> <-- caker: you should put some code in to see if an A record is an actual linode client IP, and if so, automatically add the corresponding LOC record, with either HE or TP's datacenter as the location :)
21:40<fo0bar>maybe even get it down far enough to cabinet level :)
21:40<@caker>too bad my servers aren't GPS enabled
21:40<fo0bar>so send a monkey to your cabinets with a GPS unit
21:40<fo0bar>write down the values
21:41<@caker>URGENT SUPPORT TICKET: Please provide the GPS location of my s3rV3rs
21:41<fo0bar>alt: [-100000.00 .. 42849672.95] BY .01 (altitude in meters) <-- even record the elevation of the specific host ;)
21:42<TheFirst>talk about a useless record
21:42<@caker>no way, that rocks
21:42<@caker>TheFirst: ever use geoiplookup?
21:42<TheFirst>geek points but useless imnsho
21:42<@caker>maxmind is the shit
21:42<@caker>We use it all the time to see the distance between signup IP and CC address
21:43<fo0bar>TheFirst: I could see legitimate uses for IP-enabled wildlife monitoring equipment, for example
21:43<taupehat>caker: am thinking of erikh signing up and mistakenly saying he was from Andorra
21:43<fo0bar>also, many metro busses have GPS units and data uplinks
21:44<@caker>taupehat: yeah, that's actually another fraud-score check, but yeah ...
21:44<fo0bar>remember when IPv6 came along and all of our toasters now have IP addresses? well, we need to query where our toasters are at all times
21:45<taupehat>fo0bar: my toaster is right here:
21:45<taupehat>(that's the actual machine you're connecting to now)
21:47<@caker>4h, 8h, 1w, etc better than seconds display, no?
21:47<@caker>or a combo display?
21:48<@caker>14440 (4 hours)
21:48<thoth39>Maybe that's better, as people that manage DNS are familiar with the just-integer notation.
22:03<Dreamr3>caker: how come my linode seems to slow lately?
22:04<taupehat>someone in another channel just said "badscentos" earlier today
22:05<@caker>Dreamr3: you're asking me? swap usage?
22:06<Dreamr3>caker: no, i just upgraded the memory, it's using no swap... and not really any processor... but if i leave a window open... ic ome back and it takes like 5-10 seconds for a keypress to register, then it's fast after that
22:07<Dreamr3>it feels like being swapped out though
22:34<TheFirst>caker: you'll love to know some thrasher is making it hell on 56 again
22:35<TheFirst>or better yet a kernel dump
22:38|-|^GaveUp^ [] has joined #linode
22:38<@caker>TheFirst: host looks fine
22:39<^GaveUp^>hdparm dumped my node :(
22:39<@caker> <-- mmm, pretty
22:40<@caker>slightly too neon
22:40<^GaveUp^> Timing buffered disk reads: 18 MB in 3.02 seconds = 5.95 MB/sec
22:40<^GaveUp^>not so fine :(
22:40<^GaveUp^>that black on dark gray is hard to read
22:40<@caker>you don't see the green table header background?
22:41<^GaveUp^>what do you use for the website? just flat out php or one of the mvc frameworks?
22:41<^GaveUp^>darkgray and lightgray
22:41<^GaveUp^>ah now i see the green
22:41<^GaveUp^>much more readable
22:41<@caker>^GaveUp^: no, it's coldfuson, no mvc framework other than the one I developed
22:42[~]^GaveUp^ has been fighting it for years but is afraid he has to get into webdev to have a chance at finding a job
22:42|-|thoth39 [] has quit [Quit: Leaving]
22:42<@caker>data accessors in one series of files (sql inserts, updates, etc), reusable display modules in another, functions in another set, etc
22:43<@caker>this is from many years of dealing with the WRONG way to do things
22:43|-|FireSlash [] has joined #linode
22:43<^GaveUp^>i hate the design portion of things....
22:43<^GaveUp^>ui design that is
22:44|-|TheFirst [] has quit [Ping timeout: 480 seconds]
22:44|-|^GaveUp^ changed nick to TheFirst
22:44<TheFirst>well that was so not fun
22:45|-|spr [] has joined #linode
22:57<tierra>caker: "that's the network interrupt failing to allocate memory" ... how might I go about preventing it from happening?
22:58|-|vodka [] has quit [Quit: Lost terminal]
22:58<tierra>it's never been a problem until recently
22:59|-|VS_ChanLog [] has left #linode [Rotating Logs]
22:59|-|VS_ChanLog [] has joined #linode
22:59<tierra>and I haven't changed much for some time now
23:05|-|Netsplit <-> quits: dc0e, iggy, Jay-dargo, tsi, alnr, shakr, spr, taupehat_, luis, afv-13, (+4 more, use /NETSPLIT to show all of them)
23:06|-|Battousai [] has joined #linode
23:06|-|Netsplit over, joins: spr
23:06|-|Netsplit over, joins: FireSlash
23:06|-|Netsplit over, joins: taupehat_
23:07|-|Netsplit over, joins: odedra
23:07|-|Netsplit over, joins: afv-13
23:07|-|Netsplit over, joins: iggy
23:07|-|Netsplit over, joins: dsoul
23:07|-|Netsplit over, joins: shakr
23:07|-|Netsplit over, joins: schultmc
23:07|-|Netsplit over, joins: luis
23:07|-|Netsplit over, joins: Jay-dargo
23:07|-|Netsplit over, joins: tsi
23:07|-|Netsplit over, joins: dc0e
23:07|-|Netsplit over, joins: alnr
23:10|-|davidian [] has joined #linode
23:10|-|davidian [] has left #linode []
23:38<Dreamr3>i can't seem to ssh to host22
23:40<Dreamr3>can anyone look into why load is so insanely high?
23:43<@caker>Dreamr3: which account?
23:43<Dreamr3>yyyc514, seems better now
23:43<Dreamr3>it was "frozen" for a while
23:44<Dreamr3>but the web bar was full all the way and it wouldn't respond
23:44<Dreamr3>(in the admin tool)
23:46<Dreamr3>somone using excessive IO?
---Logclosed Wed May 10 00:00:47 2006