#arpnetworks/ 2016-07-29,Fri

↑back Search ←Prev date Next date→ Show only urls(Click on time to select a line by its url)

WhoWhatWhen
brycecbrycec mutters about stupid stinking ARP tables [00:07]
mercutiodid it remember a failure? [00:07]
brycecSeems like it
(brb)
[00:07]
mercutioyou can route delete -host
usually
[00:07]
toddfroute -n get 192.168.1.1
can be revealing
[00:10]
brycec(If you're curious, pay attention to the MAC addresses http://sprunge.us/QijW ) [00:10]
toddfbut yes, 'No route to host' generally can also mean pf is blocking [00:10]
brycec(unless one has tried with pfctl -d ;))
Jeez toddf isn't it about 2am there? You're up late.
[00:10]
toddfbrycec: joys of a family and 'getting stuff done' generally means waiting till everybody else is asleep
toddf queues wife's call to bed, but is not yet done clearing kid unfriendly disaster from table
[00:12]
mercutioyou're up past midnight too, bryce :) [00:12]
brycecmercutio: by 13 minutes :P [00:13]
mercutioheh [00:13]
brycec(my fiancee just got home from work) [00:13]
mercutioi'm trying to go to bed at 9:30 pm now
means i get up really early
which seems even better than being up late for me
[00:13]
toddfgood sleep habits .. pipe dream for now .. I get up early with the older two, am up late helping mamma feed the younger one .. ;-( [00:14]
mercutiotoddf: micro nap? [00:14]
brycecwhy the hell is FreeBSD sending the ICMP response to a different MAC than it was received from?? The ARP table is correct, routing looks fine. Hmm [00:15]
toddfmercutio: that would cut into my supposed work time during the day .. otherwise I'm 100% on the go wrangling kids or other bits of life ;-( .. family needs a lifehack badly, just need to find the fuel and direction .. [00:15]
mercutiotoddf: catch 22
brycec: io clash?
ip
arping works on linux
i don't know if it's available on bsd
arping program, rather than task.
[00:16]
brycecProblem solved, was a pf rule on the FreeBSD side [00:21]
mercutiowoot
so tcpdump operates before pf
that seems kind of lame
actually for incoming traffic that'd be better
[00:21]
toddftoddf tries to imagine a world where pf_tap() is called before bpf_tap() (or whatever its actually called) [00:23]
brycecYeah I love that fact about it
"If I see it in tcpdump then I know the packet at least made it this far, and to start digging into pf" etc etc
In this case, FreeBSD side had a "(reply-to $ARP's_GW)" in the rule allowing that private traffic
Well, sorry for the noise.
[00:23]
mercutionoise is good
things have been quiet recently
[00:25]
brycec...too quiet... [00:26]
........ (idle for 37mn)
up_the_ironsup_the_irons shuts his eyes
mercutio: wow, 9:30pm *is* early
[01:03]
mercutioup_the_irons: ikr :)
it helps avoid the second wind though
[01:07]
up_the_ironsyeah, i can understand that for sure. I get my second wind around 10/11 [01:07]
mercutioyeah, so it means more balanced energy, and it means waking up really early where it's quiet like the evening
so it's kind of a win/win
[01:08]
up_the_ironssweet [01:09]
........................................................................................................................................................................................................... (idle for 16h52mn)
mercutiohttp://www.anandtech.com/show/10533/samsung-expands-its-pm1633a-lineup-as-1536-tb-ssd-hits-retail-for-10k
that seems rather high density
[18:01]
....................... (idle for 1h54mn)
mnathanimercutio: I wouldnt trust my data on an SSD with such high capacity, unless it was backed by a RAID 10 or something similar [19:56]
mercutiomnathani: i wouldn't want to spend $10,000 USD on a SSD myself...
but raid 10... that's $40k.. for 2x2...
personally i have 4 ssd in raid 10 at home
and when i get around to it i'm shifting to raidz with 3 larger ssd
there's really no iops performance benefit, and write endurance isn't really a big deal
[19:57]
........................ (idle for 1h58mn)
***Lucifer333 has joined #arpnetworks [21:56]
brycecAt that capacity and price, SSDs like that are usually destined for Big Data and Big Compute where you need more storage than you can feasibly have as RAM, that data is initially stored somewhere more traditionally reliable, spooled to SSD, then crunched.
i.e. Amazon ECS, Hadoop clusters, etc
In fact I can think of several Amazon/cloud services that would benefit from such a beast.
[21:59]

↑back Search ←Prev date Next date→ Show only urls(Click on time to select a line by its url)