Back to Home / #linode / 2006 / 12 / Prev Day | Next Day
#linode IRC Logs for 2006-12-29

---Logopened Fri Dec 29 00:00:53 2006
00:05|-|Nigel_ [] has joined #linode
00:11|-|Nigel [] has quit [Ping timeout: 480 seconds]
01:17|-|Newsome [] has quit [Quit: Linux: Now with employee pricing!]
04:43|-|LondonGuy [] has joined #linode
04:43<LondonGuy>anyone having trouble with their linodes today? I'm just experiencing some slooooowness on my host
04:43<LondonGuy>ie, I type "top", and nothing happens.. for minutes!
04:44<LondonGuy>sitting in 99% wa CPU state
04:44<LondonGuy>last time that happened another account on my host had been compromised
04:46|-|jimcooncat [] has joined #linode
04:46<LondonGuy>seems to have settled down now.
04:48|-|jimcooncat [] has quit []
04:55<warewolf>LondonGuy: go log into the website and check your host cpu graph
04:55<warewolf>see if it's high/med/low
05:02<LondonGuy>warewolf that hasn't been a good indicator for me in the past
05:03<LondonGuy>running "top" is the best one
05:03<LondonGuy>the trouble is when a host attempts too much I/O, common during a virtual server compromise
05:03<LondonGuy>there's nothing the others can do except watch the CPU sit in I/O wait state mode
05:03<LondonGuy>and occasionally get a time slice
05:04<LondonGuy>must be one of the few limitations of UML
05:04<LondonGuy>because it handles everything else rather well
05:35|-|Beirdo [] has joined #linode
05:48<Nigel_>LondonGuy, io_tokens do help prevent that
05:48<Nigel_>either way, might be an idea to submit a ticket, i think the admins are asleep (or still yelling at HE)
06:02<LondonGuy>it's ok, problem went away. I'd only raise a ticket if it was persistant and not resolving, thanks Nigel_
06:02<@linbot>New news from forums: Load spikes in the last 24 hours in General Discussion <>
06:05<LondonGuy>thanks linbot
06:45|-|ajmitch [] has quit [Ping timeout: 480 seconds]
07:15|-|ajmitch [] has joined #linode
08:15<@linbot>New news from forums: 33% Disk Space added to all Plans, Reduced Disk Space Cost in Announcements <>
08:16[~]caker moshes with you
08:23<Beirdo>I can really use that :)
08:34[~]iggy humps caker and the other linode staff members
08:35|-|Beirdo [] has quit [Remote host closed the connection]
08:37<iggy>although it says 400's are supposed to get an extra ~4G, but I only see 2
08:40|-|Beirdo [] has joined #linode
08:44<@caker>iggy: username?
08:44<Beirdo>reboot complete. Thanks muchly for the extra disk, caker
08:44<@caker>Beirdo: you got it!
08:44<Beirdo>I far prefer 70% full over 92% full :)
08:49<@caker>iggy: nm, buggie bug -- fixed
08:50[~]iggy humps caker again
08:54|-|jtgilbert [] has joined #linode
08:55|-|jtgilbert [] has left #linode []
09:39|-|gpd [~gpd@] has quit [Read error: Connection reset by peer]
09:49|-|gpd [~gpd@] has joined #linode
09:49[~]gpd does a little dance
09:49<gpd>thanks caker :D
09:54|-|Newsome [] has joined #linode
10:17<LondonGuy>hey wow 33% more disk space, you guys are great! Two upgrades in a year! First more RAM now more disk space! Thanks so much :)
10:18<LondonGuy>that reminds me
10:18<LondonGuy>I'm going to have to have a serious think about a second linode at the other datacentre soon
10:19<LondonGuy>just got too caught up on working in the UK lately
10:19<LondonGuy>I'd like to put together my own DNS slave
10:19<@caker>pretty sure we'll be utilizing a third datacenter in Atlanta in the next month or two, too
10:19<LondonGuy>Atlanta, cool caker
10:20<@caker>so East coast, Central, and West coast locations!
10:20<LondonGuy>I think I'm in the San Jose one at the moment (is it San Jose?) that part of the US anyway.
10:20<@caker>close enough (Fremont)
10:20<LondonGuy>I'll have to break out a map of the USA to see where Atlanta is.. *wanders off to google maps*
10:20<LondonGuy>whoah east coast
10:21<LondonGuy>that would be useful for, ahem, europe facing customers
10:21<LondonGuy>I'm glad to see you guys expanding
10:21<LondonGuy>I really hope that word of mouth is getting you more customers
10:23<LondonGuy>what I *really* am admiring is not only do you have a good product, but you're increasing our value for money, by upgrading existing customers when packages get better.
10:24<@caker>LondonGuy: thanks :)
10:24<@caker>Has anyone implement greylisting with sendmail, that can send me their rendered config file?
10:25<LondonGuy>hmm my message boards (phpBB) suddenly went from 8 users to 21 today.. I'm guessing a bot is creating users to spam my message boards!
10:25<@caker>LondonGuy: it sucks -- even with captcha I had bots registering users and posting spam (or just using the user's homepage for google juice)
10:26<LondonGuy>how are they bypassing captcha
10:26<@caker>LondonGuy: that's why there's htpasswd protection on the signup page on our forums (with the user/pass in the realm message)
10:26<@caker>there are scripts out there that will "read" captchas
10:26<LondonGuy>I don't think I've been .htpasswd'd (Apache Basic Auth) on linode forums.. maybe I just didn't notice
10:26<@caker>LondonGuy: click on any user's profile, or try to register a forum account
10:27<LondonGuy>I might rewrite the phpBB code to rearrange the form a little
10:27<LondonGuy>evidently it's a script relying on the status quo
10:27<@caker>I thought about that too... could work
10:27<LondonGuy>different param names, and perhaps layout the registration page a little differently
10:28<LondonGuy>can't really be bothered though, no one uses my forums anyway :)
10:28<LondonGuy>was an idea that died a natural death
10:28<@caker>I think I changed the form field names and they still got in... But, I didn't add more fields (and require them on the receiving page)
10:28<LondonGuy>"discuss your day travelling on the London Underground metro network" haha
10:28<LondonGuy> (hope I'm allowed to provide urls here)
10:28<@caker>that's fine.
10:29<LondonGuy>hmm form field names didn't stop it hey?
10:29<LondonGuy>might be looking for the text on the page
10:29<LondonGuy>time to hax0r up the field descriptions
10:29<LondonGuy>UsrnAm: Pa55wrd:
10:34<LondonGuy>so I'm curious.. how do you add more disk space to a host without rebooting it? Without bring down user accounts?
10:35<tjfontaine>hotplug and raid++
10:35<tjfontaine>and lvm
10:39|-|darkbeholder [] has quit [Ping timeout: 480 seconds]
10:45<@caker>LondonGuy: either we already had the space (bigger drives on more recent machines), or we shed/reduced the number of accounts per box
10:46<LondonGuy>ah ok
10:48|-|Beirdo [] has quit [Remote host closed the connection]
10:49|-|Beirdo [] has joined #linode
10:50|-|adamg [] has joined #linode
10:50<adamg>hi caker
10:51|-|scott [] has joined #linode
10:51<adamg>hows things?
10:53<@caker>good good -- I'll be back in 30 min
11:00|-|Kurt [] has joined #linode
11:03|-|afv-13 [] has joined #linode
11:36<afv-13>huge thanks caker
11:56|-|JavaWoman [] has quit [Quit: rebooting]
11:56|-|Kurt [] has quit [Read error: Connection reset by peer]
12:08|-|LondonGuy [] has quit [Quit: CGI:IRC 0.5.7 (2005/06/19)]
12:28|-|Kurt [] has joined #linode
13:20<Xel>caker, you did more space?!
13:24|-|adamg [] has quit [Ping timeout: 480 seconds]
13:25<Xel>Ok, reboot time
13:25|-|Xel [] has quit [Quit: .]
13:25|-|rainkid [] has joined #linode
13:26<rainkid>does all the linodes share the same net connections?
13:26<rainkid>ie, are the linode 400s on a different pipe than the linode 200s?
13:26<scott>there are two different data centers
13:27<rainkid>often times my linode gets really slow, and unresponsive, and i do not know if it is a disk trashing issue or a bandwidth issue. i was wondering if upgrading would really help...
13:28<fo0bar>they both house all different plans though, so yes, you'll be on the same network segment as other plans
13:28<rainkid>i average only 2% of the cpu usage on my linode
13:28<rainkid>and use less than 10% of my bandwith
13:28<fo0bar>rainkid: mtr can help you determine at the time if it's network-related
13:28<rainkid>so it sorta sucks that when i need a responsive machine, and my linode slows to a crawl
13:29|-|Xel [] has joined #linode
13:29<Xel>Wewt, new kernel and more space
13:29<afv-13>rainkid: are you on ratemylinode?
13:29<rainkid>fo0bar, any tips on how i set up mtr to determine that?
13:32<fo0bar>rainkid: take a base mtr reading now to a close host (network-wise, say "mtr"), see what the response times along the line are, then when you are having problems, run it again and see if the response times are higher than before
13:32<fo0bar>rainkid: mtr will also highlight areas along the way where things like packet loss happen
13:36<rainkid>ahh mtr... i mistakenly read that as mtrr :)
13:41|-|rainkid [] has quit [Quit: CGI:IRC (Ping timeout)]
13:45<@linbot>everybody walk the dinosaur
13:57<@linbot>New news from forums: FC6 Services Issue in Linux, Apache, Mysql and PHP (LAMP) Forum <>
14:18|-|adamg [] has joined #linode
14:26|-|SpaceHobo [] has joined #linode
14:46|-|adamg [] has quit [Ping timeout: 480 seconds]
15:23<Nigel_>caker, thanks for the extra space
15:24<efudd>/dev/ubd/0 9049528 8800524 249004 98% /
15:24<efudd>guess i can use it!
15:28|-|erikh [] has quit [Quit: Terminated with extreme prejudice - dircproxy 1.0.5]
15:30<Nigel_>hmmm, out of interest how much data do you reckon each SSH 'probe' uses
15:31<efudd>heh, i never uhm, used the last "free disk space".
15:32<Nigel_>efudd, 7.?
15:33<efudd>your question didn't make any sense to me so I made up an answer.
15:35<Nigel_>Each failed SSH attempt by phishers etc
15:35<Nigel_>how much data traffic roughly would that use
15:36<efudd>who cares.
15:36<efudd>In other words, do the math.
15:36<efudd>"a few bytes"
15:48|-|erikh [] has joined #linode
16:11[~]Griswald bites efudd.
16:18|-|Nigel_ changed nick to Nigel
16:18<warewolf>ugh I am so hungry
16:18<warewolf>okay time to to go fiveguys.
16:30<tjfontaine>bad pizza
17:35<fo0bar>caker: supermicro has finally entered the 20th century!
17:35<@caker>fo0bar: oh?
17:35<fo0bar>the server I bought recently has UID lights, AND a dvdrom drive!
17:36<@caker>no. way.
17:36<fo0bar>welcome to the future!
17:36<@caker>The last group of servers we bought didnt even have a floppy drive
17:36<@caker>yes, but they did havea dvd drives
17:37<fo0bar>this one did, but the bay is mutli-purpose. it can either hold a floppy or serial + USB (this one included the floppy, but the previous generation came with serial/USB)
17:37<fo0bar>unfortunately they couldn't leave well enough alone
17:38<@caker>a serial&usb port in the front bay? weird
17:38<fo0bar>while the disk bays/controller are still SATA, the cable used to connect the backplane to the mobo is multilane SAS
17:38<fo0bar> <-- so I just ordered these
17:38<fo0bar>to hook them up to the 3ware card
17:38<@caker>I'm still curious how (or why) to make the red "disk is bad" lights work with the 3ware card
17:39<fo0bar>I don't think you do
17:39<@caker>What SM model is this?
17:39<fo0bar>one sec
17:40<fo0bar>it's freaking deep compared to other models
17:40<@caker> <-- ours
17:41<@caker>non-xeon / less ram slots for your boxes, eh?
17:42<fo0bar>yeah, but we obviously don't need as much bang as you do :)
17:42<@caker>4cm 12,000 RPM Nidec dual blade fans <-- these are LOUD at full speed.
17:42<fo0bar>the 5015P-T is for a fileserver in this case
17:42<@caker>but, to be honest -- this last group of boxes seem slightly quieter than all my previous models
17:42<fo0bar>4x750GB baybe
17:43<fo0bar> <-- these are used as web servers
17:43<fo0bar>nice and cheap, standard stuff
17:44<fo0bar>thankfully the Pentium D stuff doesn't seem to care about RAM sides/banks
17:44<fo0bar>I was about to put 4x1GB in the 5015M-T and it actually shows 4GB
17:45<fo0bar>whereas the older P4-based boards were picky about ram. if you put in 4 "dual-sided" modules, it would ignore the last module (IE, 4x1GB = 3GB)
17:46<fo0bar>and of course single-rank RAM was expensive and hard to find in >512MB sticks
17:48<fo0bar>/dev/md0 2.7T 217M 2.6T 1% /mnt/wooboy
17:48<fo0bar>I just did software RAID0 to play with until those adapter cables arrive :)
18:01|-|Kurt [] has quit [Read error: Connection reset by peer]
18:02|-|scott [] has left #linode []
18:04|-|Kurt [] has joined #linode
18:04<erikh>warewolf: you should post your games on the EISOnline forums
18:07<erikh>caker: we used the serial/usbs on our machines to do quick installs with a portable drive
18:07<erikh>also, we used the serial in a couple of spots with an old modem, when the DC monkey was just wasting cash
18:17|-|darkbeholder [] has joined #linode
19:36|-|spr [] has joined #linode
19:36|-|spr [] has quit []
19:36|-|spr [] has joined #linode
20:17|-|Kurt [] has quit [Ping timeout: 480 seconds]
20:48<@caker>!seen artifex
20:48<@linbot>caker: artifex was last seen in #linode 1 year, 40 weeks, 5 days, 3 hours, 9 minutes, and 37 seconds ago: <Artifex> see you around, or whatever
22:58|-|Nigel [] has quit [Read error: Connection reset by peer]
22:59|-|VS_ChanLog [] has left #linode [Rotating Logs]
22:59|-|VS_ChanLog [] has joined #linode
---Logclosed Sat Dec 30 00:00:07 2006