Back to Home / #linode / 2015 / 08 / Prev Day | Next Day
#linode IRC Logs for 2015-08-17

---Logopened Mon Aug 17 00:00:07 2015
00:15-!-i-make-robots [~i-make-ro@] has quit []
00:36-!-ctpdump [~tcpdump@2607:5300:60:419e:5555:5555:ce5c:c64f] has quit [Ping timeout: 480 seconds]
00:45-!-RumpledElf [] has quit [Ping timeout: 480 seconds]
01:39-!-sandeep [~sandeep@] has joined #linode
01:42-!-Cromulent [] has joined #linode
01:55-!-d1g1t [~sandeep@] has joined #linode
02:02-!-sandeep [~sandeep@] has quit [Ping timeout: 480 seconds]
02:04-!-aspis [] has quit [Ping timeout: 480 seconds]
02:45-!-d1g1t [~sandeep@] has quit [Quit: d1g1t]
03:14<+linbot>New news from forum: General Discussion • Free Streaming The Gift (720p Video Quality) <> || General Discussion • Download and Streaming Fantastic Four Full Movie Online <> || General Discussion • ~Watch Online The Man from U.N.C.L.E. (720p Video Quality) <>
03:24<+linbot>New news from forum: General Discussion • ~Watch Online Trainwreck Full Movie Online <> || General Discussion • [FULL] Watch Ricki and the Flash (720p Video Quality) <> || General Discussion • Download and Streaming Minions (720p Video Quality) <> || Gener
03:28-!-getsmart [] has joined #linode
03:44-!-RumpledElf [] has joined #linode
03:49*dcraig tickles synapt around a bit with a large platy
04:07-!-gparent [] has quit [Ping timeout: 480 seconds]
04:08<hawk>Uhm... someone has found the "post new thread" button
04:09-!-aspis [] has joined #linode
04:10*dcraig tickles hawk around a bit with a large dartfish
04:11<dcraig>it's right next to the tickle button
04:35-!-JediMaster [] has joined #linode
05:02-!-RumpledElf [] has quit [Remote host closed the connection]
05:03-!-RumpledElf [] has joined #linode
05:03-!-Sputnik7 [] has quit [Quit: -=SysReset 2.53=-]
05:11-!-aspis [] has quit [Ping timeout: 480 seconds]
05:17-!-getsmart [] has quit [Remote host closed the connection]
05:23-!-getsmart [] has joined #linode
05:56<+linbot>New news from forum: Web Servers and Web App Development • How do I setup Python 2 on my LAMP? <>
06:03-!-Coobra [] has quit [Quit: leaving]
06:54-!-RumpledElf [] has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
07:15-!-aspis [] has joined #linode
07:16<+linbot>New news from forum: General Discussion • msbi Online Training in india,hyderabad,usa,uk,bangalore <>
07:17-!-sandeep [~sandeep@] has joined #linode
07:22-!-avenj [] has joined #linode
08:05-!-jottinger [] has joined #linode
08:05-!-mode/#linode [+o jottinger] by ChanServ
08:37<+linbot>New news from forum: General Discussion • Oracle 11g DBA Online training | Oracle 11g DBA Tutorial in USA,UK, Canada, Australia, Malaysia, Singapore. <>
08:39-!-jottinger [] has left #linode [Textual IRC Client:]
08:41<Cromulent>what's up with all the forum spam recently?
08:44<akerl>spammers gonna spam
08:48-!-grawity [] has quit [Ping timeout: 480 seconds]
08:48-!-grawity [] has joined #linode
08:52-!-MaliutaLap [] has joined #linode
08:59-!-jasonm [] has joined #linode
09:05-!-anomie [] has joined #linode
09:23-!-gparent [] has joined #linode
09:36-!-getsmart [] has quit [Quit: Leaving]
09:38-!-dianoga7 is now known as dianoga
09:38<sp_>hey guys
09:49<Cromulent>heh just spent ages trying to figure out why meteor wasn't registering a click event on some text and realised I had the event handler in the wrong template :|
09:49<Cromulent>stupid eyes
10:13-!-Sputnik7 [] has joined #linode
10:21-!-aspis [] has quit [Ping timeout: 480 seconds]
10:40<MajObviousman>suuure, blame the eyes
10:47<JediMaster>got a weird one, I can't change the ownership, permissions, create files or directories within one particular directory on my server that's just been created by ISP Config, using the root user
10:48<JediMaster>the ownership is root.root, permissions are 755 on the directory, yet I get: chmod: changing permissions of ‘./’: Operation not permitted
10:48<JediMaster>mkdir gives: mkdir: cannot create directory ‘backend’: Permission denied
10:48<JediMaster>I can create directories in the directory above
10:49<@jfred>try running lsattr on it, maybe it has the "immutable" attribute set
10:50<@jfred>actually hmm I'm not sure if that would affect changing unix permissions on it
10:50<JediMaster>----i--------e-- ./web5
10:51<JediMaster>strange only that and the previous site's directory have that i attribute, the older ones don't
10:52<JediMaster>heh, never come across that attribute before
10:52<JediMaster>yeah that was it, chattr -i web5 has sorted it
10:52<JediMaster>thanks jfred
10:56-!-therock247uk [] has joined #linode
10:56-!-jrhunt [] has joined #linode
11:04-!-larsdesigns [] has joined #linode
11:06<@jfred>no problem :)
11:18-!-adrian [] has joined #linode
11:48<+linbot>New news from forum: Web Servers and Web App Development • Need Help ISPCONFIG (mysql crashing) <>
11:48<MajObviousman>P.S. that little immutable trick is a common item to test for in tech interviews of sysadmins
11:50<@ccravens>I don't like testing admins for technical trivia. I want to know how they think and solve problems.
11:51<akerl>Well yes, "you're trying to write to some files but cannot, why" is a pretty nice problem solving exercise
11:53<@ccravens>If you're looking for how they approach solving the problem and it's not a trick question requiring a very specific answer.
11:53<@ccravens>If it's a question with a specific answer, that's not a problem-solving question, it's a technical trivia question.
11:54-!-getsmart [] has joined #linode
11:54<sirpengi>maybe it falls under "testing for basic knowledge"
11:55-!-jrhunt [] has quit [Quit: Lost terminal]
11:55<sirpengi>like, i wouldn't hire a javascript programmer that can't tell me what a closure is, despite it being a bit of technical trivia
11:55-!-getsmart [] has quit []
11:55-!-sandeep [~sandeep@] has quit [Quit: sandeep]
11:58<MajObviousman>ccravens: I'm also against technical trivia, but I encountered it in a tech interview, and they got to see my problem solving as a result of running into this
11:58<MajObviousman>"Why is this chmod not working?"
11:58<@ccravens>I've been a Unix/Linux systems admin for over 20 years. I've run into an issue with the immutable bit maybe twice. Knowledge of it wouldn't be a significant factor in an interview from my point of view.
11:58<@ccravens>I'm okay with it if the goal is to watch someone think out loud and see how they operate.
11:58<MajObviousman>well, what does he turn to for help first? What does he try? Did he consult a manpage? Is there a 5 minute pause in console activity indicating he went out to the web?"
11:59<MajObviousman>the only times I've run into it in the real world was when either A) I used it to confound other users on a system and keep them from fucking something up and B) When a cracker compromised a system and used it to confound me as I was fixing it
12:01<@ccravens>If it's a "working at a prompt" kind of thing, I'm more tolerant of this kind of thing. I hate being asked to recall trivia in a verbal interview.
12:01<MajObviousman>why thank you ccravens, I certainly would like some lunch
12:01<akerl>"Draw on the whiteboard what the chmod manpage looks like"
12:02*MajObviousman draws a pile of poo
12:02<akerl>No no, that's the tar manpage
12:02<akerl>no job for you
12:02<MajObviousman>the sudo manpage, which starts off describing EBNF just so you can have a prayer of grokking the rest of the manpage
12:04-!-getsmart [] has joined #linode
12:05<Eugene>Now taking bets on time to first self-righteous anti-spam complaint
12:06<akerl>I'm typing mine now
12:17-!-Cromulent [] has quit [Quit: KVIrc 4.2.0 Equilibrium]
12:22-!-avenj [] has quit [Ping timeout: 480 seconds]
12:25-!-aspis [] has joined #linode
12:28<+linbot>New news from forum: Linux Networking • Network unreachable <>
12:40-!-Cromulent [] has joined #linode
13:24-!-v0lksman [] has joined #linode
13:52-!-eagles0513875 [] has joined #linode
13:52<eagles0513875>hey guys is it possible to move a node balancer to another DC once you deploy it
13:54-!-adrian [] has quit [Ping timeout: 480 seconds]
13:55-!-swajr [] has quit [Remote host closed the connection]
13:57<eagles0513875>so i would need to decomission it and recomission it?
13:58-!-aspis [] has quit [Ping timeout: 480 seconds]
13:58<eagles0513875>ok. in a node balanced setup what is the best way to sync files between the servers?
13:58<akerl>Put your files in a CDN, the backends run your code
14:00<eagles0513875>what about a webserver for example dont you need to keep teh files synced between the two nodes?
14:01<akerl>What files?
14:03<eagles0513875>static files like html etc
14:04<akerl>Those sound like they should be in version control
14:04<eagles0513875>and they are but how would one sync them across a node balanced setup
14:04<eagles0513875>you would still need somethign to sync them between the servers no?
14:07<mcone>Lots of ways of doing that, but you could use something like this:
14:09<sirpengi>eagles0513875: put them in some version control, and then log into each server and pull from the master repo
14:09<eagles0513875>why version control and not something like rsync them between machines?
14:10-!-aspis [] has joined #linode
14:10<sirpengi>if this is code, or some project, you put them in version control *anyway*
14:10<sirpengi>if this is just a random mishmash of files, then yeah rsync would be the way to go
14:10<akerl>Because you should be tracking your project in version control anyways
14:10<eagles0513875>sirpengi: its website like a wordpress site etc
14:10<akerl>And then you deploy to the place by git pull-ing to the tag you want
14:10<sirpengi>yeah it should be version controlled
14:11<akerl>Because the goal isn't to sync between the machines, it's that all the backends in rotation have been synced based on the upstream correct version
14:12<eagles0513875>ok still a bit lost will do some more research
14:13-!-Cromulent [] has quit [Quit: KVIrc 4.2.0 Equilibrium]
14:13<akerl>eagles0513875: So you have a NodeBalancer, some backend Linodes, and then your DB cluster
14:13<akerl>All the backend app Linodes and all the DB Linodes get provisioned with your config management system of choice, and as part of that, the app boxes get a checkout of your website's repo at the desired tag/hash/branch/wahtever
14:14<eagles0513875>akerl: ok
14:14<akerl>For bonus points, have a way to pull backends who are out-of-date from rotation so you can do rolling upgrades rather than being split for a bit
14:14<eagles0513875>ok will try and draw out a diagram of how im picturing things and get yalls feedback
14:18<MajObviousman>a pretty terrible solution is to have all of your backend nodes mount a network share
14:18<MajObviousman>a only marginally less terrible solution is to use lsync
14:18<akerl>MajObviousman: I stand by my recurring statement that networked filesystems are never the answer
14:18<MajObviousman>they are an answer ... just not a good one
14:19<psandin>he did say "pretty terrible"
14:19<MajObviousman>it's somewhat dependent upon the content you're serving
14:20<psandin>if that's the game we're playing, replication over point to point IP over carrier pigeon links
14:20<MajObviousman>for example, a CMS like MODX which defaults to file-based caching and heavy file locking would perform hideously bad across NFS
14:20<akerl>locks on NFS?
14:20*akerl cries
14:21<eagles0513875>MajObviousman: wasnt suggesting using nfs
14:21<eagles0513875>was thinking of using rsync
14:22<MajObviousman>works fine as long as what's on the HD is reasonably static. But it wouldn't work well for standalone backend MODX instances UNLESS you were using IP affinity
14:22<akerl>eagles0513875: So what happens mid-rsync?
14:22<MajObviousman>because of the file caching thing
14:23<eagles0513875>akerl: in what sense
14:23<akerl>So mid-rsync, one backend is pushing files to another and isn't done yet
14:23<MajObviousman>state transition from A to B. As long as you're on either state A or state B, then you're serving files in sync
14:23<akerl>Also, where do the files come from originally, and how do the backends know who rsyncs to who? what if a backend goes down
14:23<eagles0513875>valid points indeed
14:24<MajObviousman>while in transition, you're potentially serving files from both state A and state B which may or may not agree
14:24<MajObviousman>what to do to minimize that? With enough backends and some scriptdancing to the LB, you could turn off requests to a node, rsync it, then turn it back on again
14:25<MajObviousman>there's just a lot of angles to consider
14:25<eagles0513875>with this discussion here im starting to realize that
14:25<MajObviousman>and amongst 20 experienced folk you're liable to get 12-15 distinct answers
14:26<akerl>And then there's the DB to consider
14:26<MajObviousman>honestly? The easiest thing to do is to start out simple and build it incrementally
14:26<eagles0513875>akerl: db's though you can setup with master master replication though so really doesnt that take care of itself
14:26<MajObviousman>if you're building systems programmatically (cattle, not pets), then it's not expensive to spin up a new Y to replace an existing X
14:27<akerl>eagles0513875: hahahahahahaahahahahaa
14:27*akerl collects his shattered soul off the ground
14:27-!-adrian [] has joined #linode
14:27<eagles0513875>akerl: do enlighten me please what is wrong with what i had said
14:27<rnowak>bad vibes suddenly summoned me, bad words were said
14:28<MajObviousman>mistakes were made, mostly by rnowak
14:28<akerl>eagles0513875: If I were making a list of things that can't just take care of themselves: infants, master/master DB replication, nerds
14:28<MajObviousman>in reverse order
14:30<akerl>eagles0513875: There's all the fun ways that master/master can explode, but for starters: if there's only 2 you can never have any kind of quorum, you have to decide who to write to, you have to decide who says they can be written to, if it goes sideways you have to go to the store and get vodka, if your replication goes sideways you get to choose between halting everything or moving forward
14:30<akerl>without consistency
14:30<eagles0513875>ahh gotcha
14:31<eagles0513875>i guess to start off with would be fire up a node balancer and two nodes
14:31<eagles0513875>one for website and then another for db backend
14:31<akerl>how does the NodeBalancer help with that?
14:32<MajObviousman>personaly I don't mind a Master/master (mind the capitals), as long as the I/O load is properly recognized
14:32<MajObviousman>more a fan of star topology myself
14:32<MajObviousman>oop, time for standup meeting
14:33<rnowak>I prefer mine to follow celestial bodies
14:33<eagles0513875>akerl: was trying to go for a simple setup like in the documentation i have see on node balancers
14:33<MajObviousman>orient all of your read slaves towards Neptune for fastest read times
14:34<akerl>eagles0513875: Yea, I'm just curious what the benefit of the NodeBal is if you only have 1 backend for it to balance to
14:35<eagles0513875>akerl: will then be scaling things up
14:37<MajObviousman>future autoscaling perhaps
14:37<akerl>I usually find my future workload autoscales whether I want to or not
14:37<akerl>No matter how I adjust my dotfiles, it scales up and up and up
14:37<rnowak>just adds more masters
14:39-!-daj [] has quit [Quit: leaving]
14:40-!-Ttech [] has quit [Quit: Este é o fim.]
14:54-!-Ttech [] has joined #linode
15:00<eagles0513875>just out of curiosity if one maxes out on the plans in terms of what they can scale up to can one request some custom node sizes?
15:02<jasonm>Most times you would just load balance across multiple nodes if your application required more horse power...
15:04<eagles0513875>jasonm: not horsepower that isnt the issue more like storage capacity
15:04<jasonm>Linode has a set of nodes available
15:05<jasonm>that is what you can choose from
15:05<jasonm>they don't offer custom nodes
15:05<jasonm>I wish they had more options for additional space too but it is what it is
15:07<eagles0513875>what would be interesting is if one could choose and create their own custom nodes
15:07<jasonm>well interesting and a pretty big technical hurtle to achieve
15:08<jasonm>right now with their plans they can effectively "fill" a server with nodes perfectly, without any loss of drive space, memory, or cpu
15:08<jasonm>which is much of the reason they can give you so much for $10/month ;)
15:12-!-JediMaster [] has quit [Ping timeout: 480 seconds]
15:13-!-adrian [] has quit [Ping timeout: 480 seconds]
15:15<eagles0513875>ya whats worrying me and CTO where i work is our use of couch DB as space for us can easily double as it works on a map reduce concept there for working on key value pairs
15:17-!-adrian [] has joined #linode
15:22-!-aspis [] has quit [Ping timeout: 480 seconds]
15:22-!-getsmart [] has quit [Ping timeout: 480 seconds]
15:24-!-Lever [] has joined #linode
15:31-!-getsmart [] has joined #linode
15:32<MajObviousman>they use couch extensively where I work. It makes me cry
15:36-!-adrian [] has quit [Ping timeout: 480 seconds]
15:43-!-jsager [~jsager@2600:3c03::f03c:91ff:fe56:c758] has joined #linode
15:43-!-mode/#linode [+o jsager] by ChanServ
15:50-!-getsmart [] has quit [Quit: Leaving]
15:59<akerl>eagles0513875: Um...
15:59<akerl>This is why you scale horizontally
15:59-!-jasonm [] has quit [Quit: Leaving]
16:09<praetorian>I have a vouch at work
16:09<praetorian>how does this relate?
16:09<praetorian>couch *
16:09-!-Bdragon [~bdragon@2001:470:c37f:30:f2de:f1ff:fe5a:8ef3] has quit [Ping timeout: 480 seconds]
16:30-!-Zachary_DuBois [] has quit [Quit: 平和!(HTTP/1.1 503 Zachary is unavailable.)]
16:30<MajObviousman>praetorian: is it comfy?
16:32<therock247uk>hi wondering if this is still going ahead tonight? This is a reminder that you have a migration still pending for Tuesday, August 18, 2015 at 2:00:00 AM BST (Tuesday, August 18, 2015 at 1:00:00 AM UTC). Please let us know if you have any questions! < was wondering why the ticket was closed
16:33<@caker>on your Linode's dashboard, do you see it?
16:33<@caker>Or did you do it already?
16:33<therock247uk>dont see it on it used to be there about a week ago ive done nothing on my side
16:34<therock247uk>Linode 2048 (pending upgrade)
16:34<@caker>No, it would be on the Linode's dashboard that the ticket was for (click on the affected Linode)
16:34<therock247uk>location is london
16:35<Lever>is it possible to encrypts archive header with zip tool?
16:35<therock247uk>You have a scheduled migration pending!
16:35<therock247uk>Your Linode will automatically be entered into the migration queue in 4 hours 24 minutes.
16:35<therock247uk>got it
16:35<praetorian>MajObviousman: of course
16:35<@caker>therock247uk: you can also do it sooner, if you so wish
16:36-!-Zachary_DuBois [] has joined #linode
16:41-!-Bdragon [~bdragon@2001:470:c37f:30:f2de:f1ff:fe5a:8ef3] has joined #linode
16:51-!-dealtrunk [] has joined #linode
16:51-!-dealtrunk [] has quit []
16:53-!-anomie [] has quit [Quit: Leaving]
17:07-!-Lever [] has quit [Ping timeout: 480 seconds]
17:25-!-aspis [] has joined #linode
17:38-!-larsdesigns [] has quit [Quit: Leaving]
17:41-!-Lever [] has joined #linode
17:59<+linbot>New news from forum: Feature Request/Bug Report • FreeBSD image? <> || Email/SMTP Related Forum • New linode user need help with SMTP accounts and installing webmail <> || Email/SMTP Related Forum • Citadel Imap question <> || Web Servers and Web App Devel
18:03-!-ctpdump [~tcpdump@2607:5300:60:419e:5555:5555:ce5c:c64f] has joined #linode
18:18-!-RumpledElf [] has joined #linode
18:24<Eugene>Gigabit problems: I get impatient for the 30 seconds it takes to download ISOs
18:24<Eugene>It's not long enough to switch to doing something else, but it's also too long to justify going for a snack
18:31-!-Lever [] has quit [Ping timeout: 480 seconds]
19:38<ctpdump>feels like not too long ago I was sharing a 56k modem with 10 other computers in the LAN (yes, I was privileged to have the 56k modem)
19:55-!-avenj [] has joined #linode
20:09-!-Bdragon [~bdragon@2001:470:c37f:30:f2de:f1ff:fe5a:8ef3] has quit [Ping timeout: 480 seconds]
20:10-!-aspis [] has quit [Ping timeout: 480 seconds]
20:13-!-Bdragon [~bdragon@2001:470:c37f:30:f2de:f1ff:fe5a:8ef3] has joined #linode
20:36-!-Ikaros [] has quit [Remote host closed the connection]
20:43-!-adrian [] has joined #linode
20:45-!-Cromulent [] has joined #linode
20:55-!-mgorbach [] has quit [Quit: ZNC -]
20:55-!-nisstyre [] has quit [Remote host closed the connection]
21:23-!-adrian [] has quit [Ping timeout: 480 seconds]
21:33-!-Ikaros [] has joined #linode
21:48-!-Sputnik7 [] has quit [Quit: -=SysReset 2.53=-]
21:50<+linbot>New news from forum: Web Servers and Web App Development • Need Help ISPCONFIG (mysql crashing) <>
21:57-!-adrian [] has joined #linode
22:00<+linbot>New news from forum: Web Servers and Web App Development • MySql crashing due to high I/O? <>
22:10<+linbot>New news from forum: Web Servers and Web App Development • Need Help ISPCONFIG (mysql crashing) <>
22:22-!-therock247uk [] has quit []
22:31-!-quantic [~quantic@] has quit [Quit: leaving]
22:50<+linbot>New news from forum: Web Servers and Web App Development • MySql crashing due to high I/O? <>
22:59-!-avenj [] has quit [Ping timeout: 480 seconds]
23:04-!-adrian [] has quit [Ping timeout: 480 seconds]
23:39-!-gthutch [] has joined #linode
23:39<gthutch>anyone here familiar with using youtube api on their website?
23:53-!-MikeyYeahYeah [] has joined #linode
---Logclosed Tue Aug 18 00:00:08 2015