03:48.13 | *** join/#gllug Leeds (~richardc@42-3-157-009.static.netvigator.com) |
04:55.28 | AndyMillar | mornings |
05:54.11 | *** join/#gllug Vyizis (~Vyizis@37.77.81.145) |
06:00.08 | *** join/#gllug Vyizis (~Vyizis@37.77.81.145) |
06:28.03 | ChoHag | There are 4 of us in the office of ~ 100. |
06:28.06 | ChoHag | Lazy bastards. |
06:31.13 | ChoHag | Oh now some cunt came and turned all the fucking lights on, then went and sat at the far end of the office. |
07:06.47 | AndyMillar | hah |
07:44.27 | gregj | results are back |
07:44.30 | gregj | AndyMillar: :-) |
07:44.36 | AndyMillar | good or bad? |
07:44.56 | gregj | you tell me |
07:45.25 | gregj | apparently write back is faster then write through (dah!) |
07:45.47 | AndyMillar | no surprises there :) |
07:45.52 | gregj | not sure tho how to change that using cli |
07:47.05 | gregj | however SSD caching gave about 20% improvement over raid 1+0 without ssd caching |
07:47.10 | gregj | so completly worthless |
07:47.24 | AndyMillar | MegaCli64 -LDSetProp WB -lall -aall |
07:47.46 | AndyMillar | gregj: try running the test 3-4 times on the cachecade volume |
07:47.52 | AndyMillar | you might start to see an improvement |
07:48.04 | AndyMillar | and then, you'll want to run the test just on the SSDs, incase the SSDs are the bottleneck |
07:48.20 | gregj | http://pastebin.com/Giz8x1Ew |
07:48.21 | AndyMillar | I've got Intel 320s and OCZ Vertex 4s in mine, and they're relatively fast |
07:49.14 | AndyMillar | Jobs: 3 (f=3): [mmm__] [100.0% done] [7454K/30707K /s] [1820 /7497 iops] [eta 00m:00s] |
07:49.26 | AndyMillar | can you grab that line from the outputs, or is it gone? |
07:49.39 | gregj | what is the VM4k for |
07:51.00 | gregj | on ssd one Jobs: 3 (f=3): [_mmm_] [99.8% done] [932K/4368K /s] [233 /1092 iops] [eta 00m:02s] |
07:51.59 | gregj | lastly what is the VM4k group for in the ini file ? |
07:52.41 | gregj | the blocksize. 4k is there, probably because modern drives operate using 4k sectors, right ? |
07:53.20 | gregj | I'll change number of jobs to 20, because that's more realistic of what's going to be happening here |
08:03.02 | AndyMillar | probably because we run kvm and that uses 4k blocks - at a guess |
08:03.15 | gregj | postgresql uses 4k blocks afaik too |
08:05.27 | gregj | apparently postgresql uses 8k blocks |
08:06.50 | *** join/#gllug cityLights (~nivw@bzq-218-29-26.cablep.bezeqint.net) |
08:11.44 | *** join/#gllug ess_tee_u (~NULL@host81-158-50-178.range81-158.btcentralplus.com) |
08:16.15 | AndyMillar | gregj: the next thing to try is going to be SSDs only, depending on how much you care |
08:17.50 | gregj | or trying 1+0 + ssd caching |
08:17.57 | gregj | and instead of sda+sdb, just have sda |
08:18.02 | gregj | ie, everything on the same drive |
08:18.25 | AndyMillar | tbh, you might find the SSDs aren't fast enough |
08:18.33 | gregj | lol |
08:19.11 | AndyMillar | I've had that :) |
08:19.45 | AndyMillar | what model SSDs a did you have with it? |
08:19.57 | AndyMillar | (~1k IOPS is probably 10x what you see on sda) |
08:24.38 | gregj | "IBM 256GB SATA 2.5in MLC HS Enterprise Value SSD |
08:24.43 | gregj | is what I got on invoice |
08:25.39 | AndyMillar | specced to 50k read, 7.5k write IOPS |
08:26.17 | AndyMillar | 350mbps read, 140mbps write |
08:26.33 | AndyMillar | so sequential IO is unlikely to be much better than rust |
08:26.44 | gregj | ok |
08:26.50 | gregj | tmie to do pgbench then :-) |
08:28.55 | gregj | can you test random io with fio ? |
08:29.03 | gregj | altho in general comparing it is probably not a good idea |
08:30.59 | yaMatt | morn |
08:35.59 | AndyMillar | that test is random |
08:36.30 | gregj | ok |
08:37.42 | *** join/#gllug chrisp (~wibble@host86-157-34-217.range86-157.btcentralplus.com) |
08:55.16 | mindbender | z00dax: yeah, what is he on about? :) |
08:56.14 | z00dax | not sure, he's a good guy otherwise. I blame the brighton air |
08:57.02 | mindbender | z00dax: is it expected to get told off when you help someone about linux @#centos ? |
08:58.08 | z00dax | mindbender: the regular mandate is to keep it centos specific, suse specific chatter will get the boot thrown in |
08:58.29 | z00dax | plus, you haveto give the guy his due - the problem was mostly the braind deadness of suse |
08:59.31 | z00dax | one 'interesting problem' we have is that #centos has become a place to get free consulting, free homework, and free advice on ahdoc contracting thing, whats it called |
08:59.32 | mindbender | z00dax: yeah, I could understand if there was a big chatter going on in the channel, preventing people to read stuff but once there are only leave/join messages. I wonder why that dude was in the centos channel, If I was Trevor, I'd convert him, not tell him off :) |
08:59.47 | z00dax | elance ? |
08:59.50 | z00dax | yeah |
09:00.24 | mindbender | z00dax: since you're around, got a question :) |
09:00.33 | z00dax | the flip side is that when someone comes in asking basic stuff like that : most of the regular assume he is doing paid for work, and just proxuing the questions from his client to the channel |
09:00.47 | z00dax | this theory has been proven a couple of times :/ |
09:01.23 | z00dax | sure |
09:01.26 | mindbender | z00dax: that's strange if he gets paid for not knowing how to setup basic networking but fair enough, besides we didn't know it was suse initially |
09:01.53 | z00dax | mindbender: its mostly an attitude problem in #centos.. been trying to work on fixing it |
09:01.57 | mindbender | z00dax: i'm just testing XFS on CentOS6 and It appears to be 6 times slower than the one on CentOS5, have you heard anything like that? |
09:02.23 | z00dax | no, that is strange. xfs on centos-6/x86_64 is a much better supported stack than it ever was on 5 |
09:02.24 | mindbender | z00dax: from what I've read, XFS on CentOS6 should be a lot faster |
09:02.34 | z00dax | same here |
09:03.40 | mindbender | i suppose i need to catch sandeen |
09:04.18 | z00dax | if you can define the setup, a bugs.centos.org entry would be appreciated |
09:04.32 | z00dax | the qa guys have been pretty pro-active in getting upstream traction on stuff |
09:04.56 | z00dax | also, the glusterfs guys seem to be taking up some of the xfs troubleshooting, since they recommend it |
09:05.10 | mindbender | hmm |
09:05.16 | *** join/#gllug cityLights (~nivw@bzq-218-29-26.cablep.bezeqint.net) |
09:05.33 | mindbender | yeah, it's my last week at this work place, if I can finish it before I leave, I'll report it |
09:05.54 | mindbender | will check with the people in #glusterfs |
09:13.01 | z00dax | mindbender: if you want to see a $150/hr contractor get #centos to fix his client machines, now is a great time to drop in |
09:15.02 | mindbender | z00dax: well, that's how all we started learning :p |
09:15.30 | mindbender | $150 is cheap anyway |
09:15.40 | z00dax | this guy has been around since 2006(ish?) |
09:15.46 | mindbender | lol |
09:15.52 | mindbender | well, at least he's still in the business |
09:15.57 | mindbender | what's his nick? |
09:16.07 | z00dax | Kaushal |
09:16.49 | mindbender | heh yeah just seen it on nickserv |
09:16.52 | gregj | anyone with experience in configuring bind for round robin ? |
09:17.09 | z00dax | gregj: you mean beyond just adding the A's ? |
09:17.18 | gregj | like troubleshooting |
09:17.22 | mindbender | gregj: z00dax is right :) |
09:17.41 | mindbender | gregj: you need to describe it, it's as simple as adding A records afaik |
09:17.41 | gregj | I know it is a matter of adding A's, I'm not that daft |
09:17.43 | gregj | thing is tho |
09:18.11 | gregj | some hosts using it use always same IP, others do proper round robin |
09:18.15 | gregj | and I don't know why |
09:18.22 | z00dax | dns works as a server and a client |
09:18.23 | mindbender | gregj: what's the client OS? |
09:18.30 | gregj | the only difference is that the ones that connect to single IP, are hooked to the same switch sa the dns server |
09:18.36 | gregj | it's all centos 6 64bit |
09:18.43 | z00dax | and may gethostbynames do random crap at the client side, so make sure its the right side you try to troubleshoot |
09:18.47 | gregj | client is pgbench (using libpq) |
09:18.53 | mindbender | gregj: the client is centos too? |
09:18.56 | gregj | yeah |
09:18.59 | gregj | all machines are exactly same |
09:19.15 | z00dax | do you get the right thing happen with 'dig' ? |
09:19.18 | mindbender | gregj: make sure you aren't running some sort of dns caching |
09:19.32 | gregj | I'm not |
09:19.39 | gregj | dig is showing right results (so is host) |
09:19.41 | mindbender | gregj: no avahi, nscd ? |
09:19.46 | gregj | nope, none of that |
09:19.55 | z00dax | so its libpq that is doing something wierd ( like an accidental sort ) |
09:20.23 | gregj | nope, just gethostbyname() |
09:20.29 | gregj | same version of postgresql everywhere |
09:20.33 | gregj | I'm puzzeled |
09:20.37 | mindbender | gregj: have you tried stracing it? |
09:20.47 | z00dax | yes, because you just said that gethostbyname() was doing the right thing |
09:20.48 | gregj | only variable is that one host is going through a router (on a different subnet), rest are on the same network |
09:20.59 | gregj | ones going through router work, ones on the same subnet don't |
09:21.08 | mindbender | gregj: are you a 100$/h contractor? |
09:21.11 | z00dax | plot thickens |
09:21.15 | gregj | nope :-) |
09:21.16 | mindbender | :D |
09:21.24 | gregj | more like £20ph |
09:21.32 | gregj | I wish I had 100$ph |
09:21.54 | gregj | I'm a software developer, not an admin - but I dipped my toe in that camp too - which is why I'm doing it too |
09:22.03 | mindbender | gregj: try adding it to /etc/hosts to make sure, that's the problem, i'd try that. |
09:23.20 | gregj | this has nothing to do with /etc/hosts |
09:23.27 | gregj | I can resolve the host in both cases |
09:23.39 | gregj | however, the one that is on localsubnet with the server and dns, uses only single IP |
09:23.44 | gregj | so it clearly is caching it somehow |
09:23.55 | gregj | the one that goes via router, frmo different subnet - doesn't cache it |
09:24.07 | mindbender | gregj: yeah, my point was more like, if you add the record to resolve the one you wanted it to resolve or to the one that doesn't get resolved in /etc/hosts and run your app, may give a clue if the problem is elsewhere |
09:24.41 | mindbender | gregj: just use strace to see what it actually does then |
09:25.08 | mindbender | gregj: you may as well run wireshark after that to see how it queries |
09:25.36 | gregj | yeah |
09:25.37 | gregj | thanks |
09:28.04 | AndyMillar | gregj: nscd or sssd will be caching |
09:28.15 | AndyMillar | other than that, round robin dns doesn't work as you expact |
09:28.17 | AndyMillar | expect* |
09:28.24 | AndyMillar | and can't be relied on for load balancing |
09:32.35 | gregj | that's what most people use for postgresql cluster balancing |
09:32.42 | gregj | or any other db for that matter |
09:32.52 | gregj | it seems to work fine, but only when the client is goign through a hop |
09:32.59 | gregj | which ti me is just bonkers |
09:36.36 | mindbender | gregj: not sure how a hop can be related tbh |
09:37.48 | gregj | neither do I |
09:38.21 | gregj | there are some differences in the strace tho |
09:38.31 | gregj | but they might have to do with timing, as I fired strace -f |
09:39.04 | gregj | the 'remote' one is doing few more operations |
09:39.08 | gregj | like gettimeofday(), etc |
09:39.14 | gregj | nothng that should ahve anything to do with this |
09:39.29 | mindbender | gregj: you can filter it for the network calls |
09:40.01 | gregj | true! |
09:41.17 | gregj | no differences there whatsoever |
09:41.19 | gregj | £$%T"£$^"£^ |
09:42.35 | gregj | but if I try multiple processes, it does not cache between them |
09:42.58 | gregj | I dont' know what the difference is, but it is some subtle difference in the logic somewhere, and has nothing to do directly with networking |
09:44.07 | gregj | spoke too soon, the scaling one is the remote one, not local |
09:44.08 | gregj | doh |
09:44.48 | mindbender | gregj: # strace -s 500 -e connect ping google.com |
09:45.07 | mindbender | gregj: connect(4, {sa_family=AF_FILE, path="/var/run/nscd/socket"...}, 110) = 0 |
09:45.18 | mindbender | that's what I get, maybe a similar approach? |
09:45.37 | gregj | yeah, I've seen that |
09:45.42 | gregj | that gives -ENOTFOUND |
09:45.45 | gregj | so... |
09:46.05 | mindbender | huh |
09:46.36 | mindbender | gregj: not sure why it'd say that, try -e open |
09:46.52 | gregj | strace -e open ? |
09:47.40 | gregj | exactly the same |
09:48.44 | mindbender | gregj: that can't be right, can you paste your line? |
09:49.48 | gregj | which one |
09:50.01 | gregj | when I run -e open, the nscd line doesn't appear there |
09:50.23 | gregj | connect(3, {sa_family=AF_FILE, path="/var/run/nscd/socket"}, 110) = -1 ENOENT (No such file or directory) |
09:50.43 | gregj | it tries it twice |
09:51.00 | mindbender | gregj: i think that's fine |
09:51.17 | gregj | yeah |
09:51.26 | mindbender | gregj: whatabout open() ? |
09:51.48 | gregj | funny thing is, if I run pgbench multiple times on the host that is listed in dns records -it always picks its own IP |
09:51.53 | gregj | so I think there is some funny logic there |
09:52.08 | gregj | nothing interesting there with open() |
09:52.08 | mindbender | gregj: i'm sure they didn't include their own networking stack :D |
09:52.33 | mindbender | gregj: not even open("/etc/resolv.conf", O_RDONLY) = 4 |
09:53.25 | gregj | http://pastebin.com/WtABpXFa |
09:53.39 | gregj | of course they did |
09:53.43 | gregj | but thing is |
09:53.50 | gregj | it sends a dns request |
09:53.54 | gregj | gets back bunch of records |
09:54.02 | gregj | and somehow always picks its own IP from that list |
09:54.06 | gregj | i.e. - favours it |
09:54.13 | gregj | thus breaking the RR |
09:54.53 | gregj | "The DNS protocol imposes no ordering on the resource records within a resource record set. When handling a response datagram, a DNS client (including the DNS client that forms the back-end of a proxy DNS server) is under no obligation to preserve the ordering of the resource record set as it is received. " |
09:56.05 | mindbender | gregj: that's true but on linux, it works fine (at least on centos5) |
09:56.23 | mindbender | gregj: ie. you get the other A record everytime you query something, with windows, that's not the case |
09:56.45 | gregj | thing is, bind will send different order fo A records every time |
09:57.10 | gregj | however, I think the code itself does some more there, and picks and favours certain IPs, such as - same subnet, or local ip |
09:57.17 | mindbender | gregj: you can try this |
09:57.19 | mindbender | options timeout:1 |
09:57.20 | mindbender | options rotate |
09:57.20 | mindbender | options attempts:1 |
09:57.28 | mindbender | that goes into your resolv.conf |
09:57.32 | gregj | http://www.tek-tips.com/viewthread.cfm?qid=1307952 |
09:57.56 | mindbender | but i suppose it only affects the dns server pick |
09:57.59 | mindbender | needs coffee |
09:58.22 | gregj | that isn't fixing it mindbender |
09:58.42 | gregj | I think the extra logic is in the code itself, and does shit regardless of etc/resolv.conf |
09:58.45 | gregj | time to dive into code |
09:59.04 | gregj | this is where being software dev and doign some admin stuff comes handy |
09:59.13 | gregj | frankly, I should get paid $100 ph for that |
09:59.24 | mindbender | gregj: must be using libresolv.so |
10:05.06 | gregj | AndyMillar: alternative to RR is something called pgpool2, which I tried and characterised - it doesn't scale as nicely as simple round robin |
10:09.23 | gregj | .. |
10:09.41 | gregj | <PROTECTED> |
10:09.44 | gregj | .. |
10:09.46 | gregj | no comment. |
10:10.22 | mindbender | lol |
10:10.30 | mindbender | where did you see that |
10:10.54 | gregj | it is fscking hidden like hell |
10:11.03 | gregj | whoever wrote glibc is just plain evil |
10:11.35 | mindbender | well, it's a good thought |
10:11.37 | mindbender | :D |
10:11.41 | gregj | res_hconf.c |
10:11.51 | gregj | it gets called via macro |
10:12.00 | gregj | innocently named POSTPROCESS |
10:12.04 | mindbender | how did you even get to whereabouts it is |
10:12.26 | gregj | #ifdef POSTPROCESS POSTPROCESS; |
10:12.27 | gregj | #endif |
10:12.34 | gregj | bit of tracing |
10:12.39 | gregj | and grepping |
10:13.27 | mindbender | in glibc's source? |
10:13.40 | gregj | aye |
10:14.16 | gregj | fscking urlich drepper |
10:14.56 | mindbender | fascinating |
10:15.08 | mindbender | one more thing to worry about when doing RR DNS |
10:15.08 | gregj | and of course there is no option to turn it off |
10:15.10 | gregj | £$%^£&^£^"£%^ |
10:16.07 | mindbender | why do you have one int. and one ext. network for the job? |
10:16.54 | gregj | differnet networks with VMs and servers |
10:17.09 | gregj | I had only 2 test serevrs previously, so tests had to run on a vm |
10:17.25 | gregj | now that I have more servers, I'm trying on the same subnet as the actual servers I'm testing |
10:17.32 | gregj | and ... doh! |
10:18.48 | mindbender | gregj: so, it's not a big deal then |
10:19.06 | mindbender | i was thinking who would have the DBs on 2 different networks |
10:19.18 | mindbender | at least for RR purposes |
10:19.30 | mindbender | but that'd be you. :p |
10:20.56 | gregj | <PROTECTED> |
10:20.58 | gregj | score gold: |
10:21.00 | gregj | |#define ENV_HOSTCONF>-"RESOLV_HOST_CONF" |
10:21.00 | gregj | |#define ENV_SPOOF>"RESOLV_SPOOF_CHECK" |
10:21.00 | gregj | |#define ENV_TRIM_OVERR>-"RESOLV_OVERRIDE_TRIM_DOMAINS" |
10:21.01 | gregj | |#define ENV_TRIM_ADD>-"RESOLV_ADD_TRIM_DOMAINS" |
10:21.03 | gregj | |#define ENV_MULTI>"RESOLV_MULTI" |
10:21.05 | gregj | #define ENV_REORDER>"RESOLV_REORDER" |
10:21.08 | gregj | on/off settings ! |
10:21.10 | gregj | yay |
10:22.11 | mindbender | what, you're gonna re-compile it? mad. |
10:22.31 | gregj | nope, enviroment variables ! |
10:22.50 | gregj | if they work |
10:23.06 | mindbender | ah didn't read that part lol |
10:23.25 | mindbender | that's what happens when you paste multi lines and why it's being prevented by your favorite IRC client. |
10:23.29 | mindbender | people don't read this shite. |
10:23.44 | gregj | however turning it all off doesn't ehlp |
10:23.46 | gregj | doh :( |
10:24.08 | mindbender | gregj: let me know if that ENV works |
10:24.26 | gregj | they don't |
10:24.31 | gregj | at least turning all off |
10:24.32 | gregj | doesn't |
10:24.38 | mindbender | gregj: how did you figure? |
10:24.58 | mindbender | gregj: passing the env on the same line with the binary? |
10:25.08 | gregj | I exported them one by one =off, and ran my process |
10:25.16 | gregj | export FOO=off |
10:25.22 | gregj | sets it in the env |
10:25.41 | mindbender | gregj: hmm it doesn't spawn its own shell or such i hope |
10:26.03 | gregj | that's why export works then, cos you inherit enviroment |
10:26.05 | mindbender | gregj: i'd be keen to try it with env + strace ping |
10:26.22 | ChoHag | <Boss> I'm going to add your private email address to the thing that spews the 1000s of alerts everybody ignored. |
10:26.31 | ChoHag | <Me> No you're fucking not. |
10:27.44 | gregj | thats' what we have filters for |
10:27.46 | gregj | and subfolders |
10:28.03 | gregj | mindbender: strace won't show you the extra logic in the code |
10:28.45 | ChoHag | And designated spam mailboxes. |
10:29.25 | gregj | yup |
10:29.33 | gregj | that's where all corporate shit goes in my case |
10:29.46 | gregj | all svp shuffling that happens |
10:29.51 | gregj | I couldnt' care elss |
10:30.48 | ChoHag | I have 4 boxes here. At 2300 the database machine got too busy to respond but apparently without any external cause. |
10:31.26 | ChoHag | There was a bit of extra chatter between them at the time but no reason for it. Nothing which would trigger anything at 11 and nothing in any of the logs I've found to indicate that something got triggered. |
10:31.26 | gregj | <PROTECTED> |
10:31.45 | gregj | what sort of db ? |
10:31.54 | ChoHag | mysql |
10:32.03 | gregj | yeah, it does weird shit |
10:32.11 | ChoHag | The support weenies kicked it and it was fine. |
10:35.05 | mindbender | gregj: host.conf is a sensible approach than env variables :D |
10:36.28 | gregj | yeah, but it still doesn't work however |
10:36.30 | gregj | reboot might help |
10:36.32 | gregj | will see |
10:37.22 | gregj | nope |
10:37.24 | gregj | hopeless... |
10:44.00 | mindbender | ll |
10:44.12 | mindbender | reboot it and grab a coffee |
11:20.25 | *** join/#gllug coppull (~h@181.sub-70-215-11.myvzw.com) |
11:56.22 | morsing | coppull! |
11:56.54 | coppull | ? |
11:57.04 | mindbender | coppull! |
11:57.43 | coppull | um |
12:01.50 | coppull | whats the bang for ? |
12:02.02 | coppull | im guessing im dense |
12:03.14 | Leeds | excitement... |
12:09.06 | coppull | ok here i am thinking its some kind of special irc thing lol ....yep im dense |
12:09.21 | coppull | in that case hello! |
12:10.43 | Leeds | good evening! |
12:12.07 | coppull | Days just starting here :( Morning |
12:12.41 | coppull | wish I was in the right time zone then I could get a pint |
12:18.11 | coppull | Instead its coffee and AIX punishment for me |
12:20.58 | *** join/#gllug dick_turpin (~peter@host217-34-163-30.in-addr.btopenworld.com) |
12:33.36 | *** join/#gllug cityLights (~nivw@bzq-218-29-26.cablep.bezeqint.net) |
12:38.43 | morsing | dick_turpin! |
12:38.45 | morsing | cityLights! |
12:38.57 | cityLights | yep |
12:39.08 | morsing | coppull: Where do you work? |
12:40.11 | coppull | NY state |
12:42.24 | dick_turpin | morsing: --+-O-+-- |
12:44.30 | morsing | coppull: Which company? |
12:44.52 | morsing | dick_turpin: How's Sharon? |
12:44.56 | coppull | Gov |
12:45.42 | coppull | nothing 16 floors below ground mind you |
12:47.00 | morsing | ? |
12:50.23 | dick_turpin | morsing: Working in a Saudi brothel. :-( |
12:51.13 | coppull | better hide yer booze |
12:57.42 | *** join/#gllug GNUdog (~GNUdog@sakura.voidcode.com) |
13:00.27 | morsing | GNUdog! |
13:00.29 | morsing | coppull: ? |
13:00.33 | gregj | lol |
13:13.47 | coppull | morsing: when I say below ground... usually when IT people in the US say X amount of floors below ground or in mountains they mean NSA DOD type of stuff...That I do not do ;) |
13:21.44 | morsing | ...? NFA-what? |
13:22.53 | coppull | http://www.nsa.gov/ |
13:23.49 | coppull | what im saying is i work for state gov not those guys ;) |
13:27.04 | morsing | Ok |
13:27.13 | morsing | coppull: Why GLLUG? |
13:27.53 | morsing | coppull: And also, do you know Zach? |
13:28.33 | coppull | Linux fan and im a Brit |
13:28.45 | coppull | I do not know Zach |
13:38.55 | *** join/#gllug funkyHat (~m@funkyhat.org) |
13:39.50 | dick_turpin | Are you Brit Eklnad? |
14:00.40 | *** join/#gllug funkyHat (~m@funkyhat.org) |
14:10.37 | *** join/#gllug funkyHat (~m@funkyhat.org) |
14:16.50 | morsing | funkyHat: STOP IT! |
14:17.40 | funkyHat | hrm |
14:18.30 | funkyHat | Not sure what's going on |
14:19.06 | funkyHat | Doesn't seem to be my usual irssi failure mode either (where my silly scripts manage to start 2 instances of it which then battle for my nick) |
14:48.13 | morsing | funkyHat: Erm... Do you like doing things the hard way? A bit of S&M? |
14:50.19 | funkyHat | morsing: mostly it's nice and good and easy (log in to special user account, irssi screen session automatically attaches or is spawned if it wasn't running) |
14:51.10 | funkyHat | I think today my issue may be transient network problems on the VPS host, though, rather than problems with my stuff |
14:51.36 | funkyHat | Or were you referring to irssi in general? ;P |
14:54.25 | toxboi | Did anyone experienced power cut in Central London like 15 minutes back? |
14:56.04 | yaMatt | not here, what part? |
14:56.24 | toxboi | Old Street/Moorgate |
15:15.15 | morsing | funkyHat: I am asking because I was reading about your Apache -> lighthhtpd |
15:16.04 | funkyHat | morsing: oh right! |
15:17.00 | funkyHat | morsing: lighttpd's config is much cleaner than apache's, and my drupal config generation thing is really handy |
15:17.12 | *** join/#gllug chrisp (~wibble@host86-180-144-97.range86-180.btcentralplus.com) |
15:17.18 | morsing | chrisp! |
15:19.53 | funkyHat | Just reloading lighttpd's config sets up new drupal vhosts based on which directories are in the sites directory |
15:39.52 | *** join/#gllug coppull (~h@181.sub-70-215-11.myvzw.com) |
15:41.40 | chrisp | morsing af'noon |
15:43.02 | yaMatt | 'noon |
16:12.35 | *** join/#gllug {mikey} (~mikey@146.185.18.114) |
16:13.01 | *** part/#gllug dick_turpin (~peter@host217-34-163-30.in-addr.btopenworld.com) |
16:13.42 | {mikey} | hi |
16:16.26 | {mikey} | I've just uploaded a script to my blog which I find extremely useful with my day to day Linux life if anyone's interested |
16:27.00 | {mikey} | http://www.green-spot.co.uk/2013/05/21/storing-and-searching-code-snippets-with-metadata/ |
17:52.47 | *** join/#gllug Armand (~martin@87-194-165-154.bethere.co.uk) |
18:08.43 | *** join/#gllug funkyHat (~m@funkyhat.org) |
18:29.43 | *** join/#gllug funkyHat (~m@funkyhat.org) |
21:33.09 | *** join/#gllug {mikey} (~mikey@cpc3-dals17-2-0-cust54.hari.cable.virginmedia.com) |
23:49.05 | *** join/#gllug Leeds (~richardc@42-3-157-009.static.netvigator.com) |