[01:11] *** m0unds_ has quit IRC (Ping timeout: 246 seconds) [01:11] *** m0unds_ has joined #arpnetworks [01:20] *** RandalSchwartz has quit IRC (Ping timeout: 256 seconds) [02:12] woot, got a 10x improvement in guest to host network performance [03:38] *** qbit has quit IRC (Ping timeout: 272 seconds) [03:40] *** qbit has joined #arpnetworks [03:41] *** qbit is now known as Guest5445 [04:10] http://man7.org/linux/man-pages/man7/lvmcache.7.html [04:10] has anyone seen this before? [06:40] *** Guest5445 is now known as qbit [19:23] acf_: i am trying to access your smokeping. did you disable it for external access? [19:25] hmmm.. [19:25] also your / exposes more than it should, but loads. [20:05] *** medum has quit IRC (Ping timeout: 264 seconds) [20:07] *** medum has joined #arpnetworks [20:51] seems there's some congestion tonight [20:54] yeh [20:54] on level3 [20:54] can you do a mtr? [20:55] yea, it's the return path [20:55] i'm trying to track it down, but i reckon it's level3 [20:55] it is [20:55] it's really bad to my uk server. [20:55] it's ugly [20:55] haha [20:55] i'm seeing 60ms +/- 150 [20:55] oh i just saw loss, and throughput issues [20:55] udp loss was higher than icmp loss [20:56] but latency seemed normal [20:56] but throughput seemed even worse than the udp loss [20:56] i just got a vultr vm that's on level3 [20:56] there's only one location in the US that routes via level3 from arp, most are ntt [20:57] yeah, udp loss is gross [20:57] what's your ip? [20:57] like 50+% [20:57] arp IP? [20:57] oh wow i didn't see it that bad before [20:57] nah ip you're coming from [20:57] 73.26.90.138 [20:58] http://pastebin.com/HUZS2VD0 [20:58] yeah it loosk strange [20:59] not as bad that run [20:59] it's up and down a lot [20:59] yeah [20:59] that's with -u? [20:59] yeah [20:59] on the bottom, forgot to redo the top [20:59] yeah do icmp [21:00] it's bloody hard to read with all that hop alternating [21:00] it seems you're not getting loss just latency from that though [21:00] and from my mtr from arp to you it's the same [21:00] it actually looks like comcast issue in this instance :/ [21:00] looked largely the same [21:01] yeah just trying to track down wehre it's happening [21:01] try doing 500 of them [21:01] i realise it takes nearly 10 minutes :) [21:01] but if there's a couple of percent loss or something it shows it up better [21:02] well, there's nearly always loss at any time of the day on level3 w/icmp [21:02] yeah on those earlier hops [21:02] that's been my experience w/them forever [21:02] it's whether it carries through [21:02] i don't tend to have any loss at the destination, just along the path [21:03] but ssh is laggy? [21:03] but when i mentioned congestion, i saw a lot of loss. it seems to just be high latency now with no loss [21:03] it isn't [21:03] That's what she said!! [21:03] BryceBot: no [21:03] Oh, okay... I'm sorry. 'it isn't' [21:03] i'm getting good speeds from this level3 connecting vultr vm :( [21:03] ahh ok [21:03] ssh wasn't laggy it seemed to uk for my "cl" program [21:03] but i was getting really bad throughput [21:03] starting around 6 hours ago [21:04] i updated my kernels and got some speed boost in general [21:04] http://uk.meh.net.nz/cgi-bin/smokeping.cgi?target=Curl.chi [21:04] i should fix the naming [21:04] that's la.meh.net.nz [21:05] http://uk.meh.net.nz/cgi-bin/smokeping.cgi?target=meh.lameh [21:05] but ping has been ok [21:05] and loss isn't that high [21:06] http://pastebin.com/BNiUsqcP [21:06] comcast's routing is changing a lot grr :/ [21:07] what throughput do you get atm? [21:07] it seems it's going via los angeles or denver [21:08] yea, it changes depending on time of day [21:08] sometimes it's direct abq > lax [21:08] it changes within the trace [21:08] sometimes it's denver > lax [21:08] but that may be the jitter [21:08] route change adds 10ms to ping but that's it [21:08] throughput is nearly always the same [21:08] and it's still ok? [21:08] yep [21:09] hmm damnit [21:09] so that's different to the other issue :/ [21:09] actually nope [21:09] it just tanked [21:09] was going 1.2MB/sec, dropped to 400KB/sec [21:09] that's pretty slow both times [21:10] yeah that's more like wha ti been seeing [21:10] except lower because it's from uk [21:10] well, here [21:10] higher ping makes loss worse [21:10] http://pastebin.com/3HhgkHDa [21:10] two local hops from arp > comcast are gross [21:10] http://pastebin.com/9FhM7qHX [21:11] yeah, that same 150-200ms "worst" measure is in the same spot as mine [21:11] 5 & 6 [21:11] yaeh hard to know if that's deprirotisiation or not [21:12] the latency to hop 5 for me is going hmm [21:13] bi'm still running my download heh [21:13] average speed 49305 bytes, current speed 45476 [21:13] it was going 80k/sec before [21:14] anyway why's it fine with this vultr vm grr [21:14] by fine i mean 7.6mb/sec [21:15] oh shit [21:15] ? [21:15] it's coming in from vultr via nlayer [21:15] i wonder if that has any bearing [21:15] oh and now it's showing loss on last hop [21:15] and on mzima pings heh [21:16] did you stop your dl? [21:16] nope [21:16] dl slowded down [21:16] its' getting slower nad slower [21:16] funky [21:16] 27k/sec now [21:16] wow [21:16] yeah i know it's weirdly slow [21:17] and how does 6% icmp loss fit in with that? [21:17] with 7.6mb/sec [21:20] damnit i have to wait for dns ttl :( [21:24] this is making no sense [21:24] so there's packet loss from la to miami [21:24] but throughput is good [21:25] and the packet loss starts at gtt in san jose [21:25] and throughput is good in both directions [21:28] maye it's bursty [21:37] heh " curl -q http://weallsee.net/testspeed | sh [21:37] " [21:37] is convenient [21:37] but it feels dirty just running stuff downloaded off the internet [21:37] it could be anyhting [22:00] m0unds: i'm setting high jitter to you from jp.meh.net.nz over ntt too [22:00] so it may be your path to arp [22:01] or within comcast's netowrk [22:01] i'm loading my connection up quite a bit [22:01] atm [22:01] ahh ok [22:01] maybe that's it then [22:02] i'm trying to get my systems better organised for pushing smokeping etc :) [22:08] mercutio: what os are most of your vultr vps? [22:09] ubuntu [22:09] one is arch [22:09] if there was arch auto provisioning they'd all be arch [22:09] i've just shifted them all to custom kernel though [22:09] any performance benefits? [22:09] a little [22:09] to the kernel [22:10] i'm not sure why :) [22:10] i did the kernel when i was trying to debug guest-host kvm performance at home [22:10] Startup finished in 944ms (kernel) + 491ms (userspace) = 1.435s [22:10] that's sj [22:11] Startup finished in 948ms (kernel) + 732ms (userspace) = 1.681s [22:11] and that's my home vm [22:11] home vm is ubuntu [22:14] and tcp performance went up a little [23:26] mercutio: hmm, haven't touched the smokeping config [23:26] probably just congestion or something [23:26] huh. it's acting funny for me too [23:28] [....] Starting latency logger daemon: smokeping [23:28] >:( [23:28] all of the smokeping things are blocking on something [23:33] huh. that was really weird [23:33] seems that smokeping was blocking on DNS resolution [23:33] both the CGI and the service [23:36] damn [23:45] i wonder why [23:55] oh it didnt' seem to lose data? [23:55] cool 4.2.2.1 is screwed [23:56] from about 2pm your time [23:56] aare you pdt? [23:58] hmm and phoenix.acfsys.net [23:58] is that a box you can trace from? [23:59] oh it's a tunnel or something