oh super
my vultr vps is down :-(  Packets: Sent = 60, Received = 3, Lost = 57 (95% loss)
what location?
vultr's network can be kind of bad
New Jersey
There is a problem with the machine your VM is hosted on. Our staff are currently investigating, and will have the issue resolved ASAP.
I guess you get what you pay for
it was down for ages in japan a while back
you have to ask if you want them to credit your account for downtime.
 i'm only using vultr for smokeping/bw testing though
and to have a reference point etc.
but dallas and uk have hd a few issues recently too
and sydney was screwed for ages
I am wondering if Digial Ocean is any better?
i dunno
s/Digial/Digital
<mnathani_> I am wondering if Digital Ocean is any better?
I've got my $5/month DO fbsd building 8.4 and 9.3 packages now.
so when I move my ARP machine to 9.3, I'll already have all the packages built.
looks like 9.3 is now EOLed at the end of 2016... that'll be nice.
I was worried about moving all the way to 10 quickly
yeah, i kind of regret going 10.1 with this box
what was the biggest problems?
a few ports were incremented that caused issues for a couple webapps i use
more annoying/inconvenient than anything
I, on the other hand, am quite happy with my FreeBSD 10 systems
(fwiw)
generally, i'm fine with it
9.x -> 10 was awful as an upgrade. clean install would likely have been a lot less of a headache for me
I don't remember having any significant issues 9->10
i upgraded pretty quickly after 10 came out to try it out, and it seemed like they had reports of the issues w/some of the upgrade process and fixed them (namely, new paths not being created)
so tons of stuff would break at boot
heh
i can't believe how broken vpn support is on android 4.4.x
so, I got the goahead from the client to develop in angularjs using the google materials plugins.  neat.
seeing intermittent packet loss via ipv4 again
argh
argv
argc
**kwargs
**kwargs, lol
so is there going to be a freebsd 9.4?
Don't think so considering 9.3 is considered legacy
(and it's not listed on releng, just 10.2 and 11.0)
10.2?
https://www.freebsd.org/releng/index.html November 2015
oh it's like 8 months out
zx
cv
up_the_irons: out of curiosity, did you end up locating the customer's switchport
with the 1Gbps spikes?
sweet freebsd developer is looking into freebsd 10.1 hang issue now (and can reproduce)
acf_: not yet, i'm pretty perplexed.  i don't see those spikes today though.  there was some traffic around 200 Mbps, but that's about it
the last loss on my smokeping was about 15 hours ago
but m0unds said he was seeing loss 6 hours ago
maybe unrelated.
yep, from 0900MDT -> 1100MDT
yeah
it would go from ipv6 to ipv4
it was weird, but it's done it two days in a row
similar timeframe too
just didn't have the opportunity to traceroute or anything yesterday because i was flying home
it wasn't as bad today as it was yesterday though. it triggered multiple svc monitoring alerts yesterday
maybe it's some new exploit
i dunno what was with gtt the other day
oh www.gtt.net isn't showing cloudfare anti-ddos page today.
> 08:43 < acf_> oh super
last loss I saw
i saw packet loss at 1316mdt last time, but it was short lived
other stuff was ~5-10mins' worth
http://kremvax.acfsys.net/smokeping.cgi?epoch_start=1425884400;hierarchy=;epoch_end=1425895200;target=Remote.googledns;displaymode=n;start=2015-03-09%2000%3A00;end=2015-03-09%2003%3A00;Generate%21=Generate%21
there it is^
is there some way to cut/paste long urls with weechat/tmux?
multiline ones have the user list next to them, so you have to cut and paste 2 or 3 lines
oh wow that's severe
but 4.2.2.1 is fine
I use irssi
the URL just wraps, and xfce4-terminal detects it
hmm
i'm using urxvt
hah i can press ctrl-down a lot and make the text tiny and copy
8.8.8.8 outbound path at least is routed via any2ix
i dunno about incoming path though, and google don't have looking glasses
A port of Codel exists for BSD and is available in pfsense and elsewhere.
i didn't realise fq_codel had been ported to freebsd
Weechat 1.1 has a keybinding for re-wrapping URLs so they're easier to select or click on.
alt+l by default.
there's a weechat script you can use to call out to bitly api or whatever too
I run that script on my work machine; my only complaint with it is that it runs in the same thread as weechat itself, so if bitly is slow, weechat is hung.
yeah, i've noticed the same thing
alt-l seems to do nothing for me
i have 1.1.1 apparently
oh i have 1.0.1 running
it's been upgraded but i haven't restarted.  i had to version myself :)
nice
maybe i should restart sometime
^ that's the command. It toggles, apparently.
but it's only in 1.1?
I believe so.
Feel free to try it in 1.0.1!
i can't remember how many times i tried alt-l :)
i can't see it in /key, but it's really long
It's just between meta-k and meta-m !
oh yeah it's not there
ther's also meta2 and meta-j-blah
226 key bindings for context "default":
bah i'll try it
wow
that was actually pretty painless
I use urlserver, so rather than weechat depending on an external API, it serves me "shortened" URLs directly :)
it even saved all my session state :)
ok now i need another long url :)
Also, alt-l has been in weechat for as long as I can remember, so... several versions. Works in 1.0.1
mercutio: like this https://wiki.opnsense.org/index.php/OPNsense:So_why_did_we_fork%3F
that's not long haha
i still don't have alt-l i think!
sorry, it's the longest url I have open :P
Maybe you mapped over it, mercutio ?
nah it's just unmapped
https://weechat.org/files/doc/weechat_faq.en.html#terminal_copy_paste
just need to find what it maps to
oh!
ctrl-alt select works
Oh also http://dev.weechat.org/post/2014/02/16/Bare-display
or nick list at top :)
i just turn the nicklist off altogether
oh nick list at help still shows some other stuff
I'm guessing mercutio your config pre-dates its addition, so it was never added/bound
yeah it's not there helpful other than knowing how many people are around in a split
i see
"use /key missing to add the key"
oh that works
it even loses colour
http://img.phluxbox.com/screenshots/ou2Aeh.png
no nicklist for me :)
 meta2-23;3
i wonder what key that is
that's about what mine looks like phlux
just without the fancy coloration
♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥
not bad, phlux
i have timestamps
Also, avoid ownCloud at all costs.
timestamps are actually quite handy
I toggle my timestamps
ahh
yeah, same here
my timestamps are boring
I always color them, but I lose the color for some reason :\
Just boring ol' timestamps
every now and then i like weechat more
I'm using weechat
that /upgrade -quit then doing weechat --upgrade  was nice
(clearly)
http://img.phluxbox.com/screenshots/gei6Og.png there's a shot of the whole wm
i enabled timestamps for that one :P
hah i'm #zfsonlinux and the nick list takes up the whole screen
I don't see the point of a nicklist in a cli client tbh
How did I not know about /upgrade... that shit's awesome.
it's just to know big splits are
I use /upgrade often
mercutio: I can see that by the nick count in the status bar :P
I agree, nicklist is kinda pointless... but I still show mine, just habit.
I'm on -git
oh
I probably need to upgrade..
hah zfsonlinux has 295
WeeChat 1.1-rc1 (git: v1.1-rc1-2-g966a7b5) [compiled on Dec 25 2014 00:16:01]
As you can probably tell in my full screenshot, I don't really have enough real estate on this laptop for a nicklist anyways
 /bar hide nicklist
oops
i figured you had low res by your font
aye
i don't think i have anything installed to take a screenshot
scrot ftw
i have import
Good ol ImageMagick
https://weallsee.net/screen.png
so the side nick thing wasn't taking up too much space
It's always surprising when a screenshot fills my entire screen
this is after goign to lower resolution
the notion title bars need to be made smaller again :/
you monster
(for having such a hi-res display)
it's only 1440p
"only" when poor phlux has < 1080p...
he's got a laptop
the keyboards on laptops are bad enough :)
I like my ThinkPad's keyboard
i use mechanical keyboards.
as do i
even on spare machines.
doesn't mean I can't also like non-mech
yeh.
my hp laptop has terrible keyboard.
at least it has the non trackpad mouse thingy
i found 1440p to 4k really didn't help /that/ much
like you can onyl focus on so many things at once...
and you can only have text so small
aye, this is a ThinkPad
I like its keyboard also
I just got a Filco Majestouch Minila.
brown switches.
i like brown
i have brown, blue, red on dfif keyboards.
I installed o-rings.
i make so many typos with red if i'm not careful.
i really should swap back to brown :/
i have o-rings too.
even the thin o-rings were good enough for me, and i split them across two keyboards...
it takes a while to pop all the keys off
switches..?
I just got cheap o-rings.
$2 for a pack of 100.
i got the wasd ones.
i don't think it was terribly expensive
Probably like $10-$20.
i got them via ebay cos it's easier with paypal etc
yeh something like that
a bit more than that
The minila is a tenkeyless keyboard. It took me about 30 minutes to install the o-rings.
$17.99+$6.55 postage.
my brown keyboard is 10keyless
it was really good after i bound mouse button to enter key
I got mine on Amazon for about $2 with free shipping.
i didn't realise that i kept on pressing enter with my left thumb from mouse
i don't need numeric pad - but that enter key is really handy when browsing etc.
so i just rebound back or forward to it
caze: damn that's cheap :)
ebay i think had some from china pretty cheap.
amazon is pretty expensive for international shipping
it only really works well for fullfilled by amazon stuff too
which is most stuff.  except motherboards.
http://www.amazon.com/008-Buna-N-O-Ring-Durometer-Black/dp/B000FMWLR8/ref=sr_1_2?ie=UTF8&qid=1425949938&sr=8-2&keywords=buna+o-ring
Amazon: "008 Buna-N O-Ring, 70A Durometer, Black, 3/16" ID, 5/16" OD, 1/16" Width (Pack of 100)"
This item does not ship to Auckland, New Zealand. Please check other sellers who may ship internationally. Learn more
yeah i can't get those ones
http://www.ebay.com/itm/Sales-150pcs-White-Rubber-O-Ring-Switch-Dampeners-For-Cherry-MX-Replace-Part-IS-/271463970610?pt=LH_DefaultDomain_15&hash=item3f34837732
can get those though
do you like having o-rings?
Yeah, it's a lot quieter.
i was thinking about feel wise.
Ideally, you don't notice them, because you're not supposed to bottom out the key in the first place.
They're like car bumpers; ideally, you never feel them.
yeah
i press quite hard so i bottom out a bit with red.
but when i flow better i stop bottoming out so much
brycec, sweet
something tgey must of changed from moanibg mertkes like me :-D
i hate this keyboard
i've never found a keyboard i really loved.
i'm a lot happier with current mechanical keyboards - but i still feel they could be beter.
but laptop keyboards are just varying degrees of hate
curiously i found i can plug a mechanical keyboard into my cellphone even :)
I'd like to try Matias quiet click switches.
noise isn't a real concern for me - feedback is.
I've plugged my mouse into my phone.
tbh, i think this keyboard probably needs a clean, some of the keys press easier than others.
i just have a cheap phone, that's only 720p, but it has lte.  and ssh from it with a real keyboard is quite comfortable.
s/and/but/
<mercutio> i just have a cheap phone, that's only 720p, but it has lte.  but ssh from it with a real keyboard is quite comfortable.
If I had time or the inclination to assemble a keyboard, I'd get one of the ones that pops up on massdrop.
i mean it's a tiny screen - but lte makes ping quite bearable.  and difficult keys are much easier to press with a real keyboard.
https://www.massdrop.com/buy/infinity-keyboard-kit
Something like that should be really easy to clean.
i have some coolermaster mechanical w/red switches and no 10key
it's been ok, not too noisy but i don't really bottom out the keys
the key markings are lasered on the front of the keys vs the top, so you can't see them unless you look at an angle. kind of snazzy.
i stopped looking at keys when i started typing in dvorak
i often type in qwerty still, but the habit died.
is it normal for my nginx to be missing 'sites-available' directory
i've only seen it on ubuntu (might also be on debian, but i wouldn't know)
it's there on Debian
mnathani_: i don't think arch has that
it doesn't
ubuntu just copies debian packages :)
my nginx on arch has sites-available
but it's empty,
i wasn't sure about debian having it because i don't use it, i assumed that's probably why it was there
i probably copied config from ubuntu
yea
i just end up using that scheme because i'm used to it from years of ubuntu use
debian has gone to that with lots of things.  i hate the new default unbound config.
they removed all the documentation from the config file, and instead just had a really cut down config
at least there's unbound.conf.example
I think it depends on the package maintainer
That's what she said!!
is it?
package maintainers have a lot of room to do what they like I think
yeah lots of packages maintainers screw around a lot
sometimes you read about developers being annoyed at it
at least i did in the past
i've noticed that lots of more recent packages seem to be maintaining themselves.
"maintaining themselves"?
ie the git repository will have debian/ etc directories.
oh I see
I suppose that's good
and they'll be more tied in with upstream pushes.
well it's better than someone who has no idea how the code works pushing bug fixes etc.
i mean when it first came about a whole lot of people just had tarballs for their software etc.
and sometimes fixes were needed for compiling
but now days if you write code and it doesn't "just work" on ubuntu you have issues.
a lot of the larger projects have build bots and incremental changes get compiled/tested on different os's too
yea
for multiple sites the sites-available/sites-enabled scheme does work pretty well
I use it
for smaller configs it doesn't matter either way
mercutio: you use OpenBSD a lot yeah?
I used to use it a lot, but then I got used to dpkg / apt-get
not a lot, but a bit
i use arch and ubuntu more.
but i'm familiar enough with openbsd.
and freebsd.
Debian dropped SPARC support recently
which is annoying for me
damn.
openbsd's sparc support has always been pretty good
yeah. so I'm trying OpenBSD and NetBSD on them
*the sparc boxen
but I'm trying to figure things out like how to manage software upgrades effectively
the last time i tried a sparc it was really slow generating ssh keys
and showing text to the console
software upgrades are a bit of a pita with openbsd tbh
there's a new openbsd binary upgrade thing that i haven't looked at yet
yea, I have two sparc boxen
one runs at 600 MHz
and is currently running my IRC client
did you max out ram?
ram should be cheap now
I think it has 512 MB
not sure about the max
oh not too bad then.
i hate the sparc ethernet interfaces
at least the older ones.
this one uses the 'gem' driver
yeah that's one of the not so great ones
I think I had trouble getting IPv6 to work over it
i think hme was the other not so great one
apparently the current openbsd hackathon is about smp
probably fixing locks in various ways ..
openbsd has most things in one big lock
NetBSD put a big focus on getting rid of the giant lock a while ago
yea
netbsd is pretty nice, i like it a lot
I tried building gccgo on it yesterday
but the build failed :(
doh
yeah openbsd has been a bit behind on that side
at least they're doing lots of positive software development
yeah, netbsd and openbsd both have clear strengths in different areas
i used netbsd as my primary server os for personal projects for a year or so
npf is pretty cool, but sparsely documented
pkgin works really well as far as binary pkg mgmt goes
I remember pkgin being really great, almost like dpkg
but it doesn't seem to manage the core system
right, it's like pkgin
err
pkgng
just handles packages you install on the base system
openbsd is ilke that too
the base system doesn't have packages
for openbsd, you use pkg_add et al, right?
you can use sysupgrade to help make base system updates a bit easier
but it's still kind of a pain in the ass, despite any automation their extra tools provide
usually with netbsd I end up running CURRENT
sometimes the release will be really unstable, and I'll try CURRENT, and it will work great :P
heh
yeah, that was the case w/my crappy little netbook
yeah for openbsd there's pkg_add
which works pretty well tbh
does it have an automatic package upgrade thing?
pkg_add -uvr
-u upgrade, -v verbose, -r recursive
err -r is replace actually
does that just go through all of your installed packages and upgrade them?
yeah
you need PKG_PATH to point to the mirror
arp has a mirror :)
nice
PKG_PATH=http://mirrors.arpnetworks.com/pub/OpenBSD/snapshots/packages/amd64
so something ilke that for -current
ugh, awful network blip
and s/snapshots/5.6/  for 5.6
m0unds: damn you still having issues
mtr time!
it stopped again
m0unds: just leave a mtr running in tmux?
that's what i normally do if there are issues that come and go
m0unds: how do you see it? just notice ssh slows down?
the only problem is that i can't get connected to the box at all when it takes a hit like that
keystrokes stop registering
can't ping the vm
ah
ssh sessions drop
m0unds: but if you connect to it when you can connect to it, and run tmux, and start a mtr to your hoem connection you can get to it later
I've been seeing ~50%, not 100%
what kvr are you on?
3 i think
3 has been having issues.
has it?
yeah.
yea, i'm on 3
yeah it seems isolated to 3.
so yeah mtr won't help, and yeh, it pauses for ages.
ssh should recover, but it can take quite a while
yep, it's super annoying
it's dropping pretty hard rather than being really jerky.
up_the irons mentioned s8.lax recently
how long has it been doing that?
m0unds: a couple of days
ah
it's different from the other issue that hits more people, it's more frequent too :(
he said that the 1 Gbps spikes came dedicated customer was connected to s8.lax
ah
is that a new switch or something?
it's the new router
s1 is the old router which most people are connected to.
ah really! I hadn't heard about it
what model?
cisco
i actually can't remember
so is he trying to phase out s7.lax
the pos
i'm sure it was discussed sometime.
ohhh
hangon
i got confused between s7/s8
s8 is a switch i think
i don't think s7 is being phased out!
just grepped my logs: 2014-08-19 01:19:20        @up_the_irons   acf_: yes, Metal customers get 1 Gbps to s1.lax, through a pair of Foundry switches (s8.lax and s9.lax).
ah, makes sense, thanks
sho 'nuff
i wonder if cddp is on
I disagree, upgrading OpenBSD is really, really simple and easy. Boot to the new ISO/ramdisk, or if you're daring you can unpack the updates in a live system. Once unpacked, reboot to the new kernel/binaries, sysmerge to update configs, and pkg_add -ui Then again, I guess having done this two-dozen times in the last month it seems easy... but I think it's pretty objectively simple.
Also, the new "right" way is /etc/pkg.conf's installbase=
brycec: i'm talking about for fixes betwen releases.
rather than defining PKG_PATH environment
oh i didn't know about that installbase thing brycec
(I also set ntogo=yes)
mercutio: the errata stuff? that's easy too. Especially with mtier
*mtier's "openup" tool
well there's stp, but that doesn't tell me what switch plugged into
(M:tier package up the errata updates and it's just a pkg_add -u then.)
s/dd/d/
<brycec> (M:tier package up the errata updates and it's just a pkg_ad -u then.)
brycec: i really should look at mtier
You really should :P
oh nice.
http://www.mtier.org/solutions/apps/openup/
does it support -current?
No
But that would be silly
*cough*
i have old systems, and -current system
If you're running -current, you just install the latest snaps.
yeah i do sometimes.
i don't have muuch open anyway
There's no errata for -current
oh, that's neat (mtier openup)
yeah i thought it may update binaries, but actually the kernel stuff complicates that with openbsd.
binpatch ^
@google openbsd binpatch
103 total results returned for 'openbsd binpatch', here's 3
M:Tier - Stable packages and binpatches (https://stable.mtier.org/) Introduction. Keeping your installed OpenBSD packages up to date is hard and  time-consuming. Nobody wants to read the mailing lists to spot security fixes ...
The binpatch-ng framework (http://opensource.mtier.org/binpatchng-README.html) It is derived from the original binpatch for OpenBSD as released in Version 1.0 by  Gerardo Santana and has since been maintained and further developed by ...
Binary patches for OpenBSD (http://openbsdbinpatch.sourceforge.net/) The binpatch framework resemblances the OpenBSD ports subsystem in many  ways. It's no coincidence since binpatch took ...
heh
i wonder if em1 or p4p1 is primary ethernet
em* numbering starts at 0, so probably not em1
i really don't understand ubuntu's network names.
true
well i'm using em1 as primary
so i suppose that means i'm on second switch
mercutio: it's physically descriptive, eg, enp8s0 is ethernet, pci slot 4, slot 0
brycec: why's em1 em1?
it's even weirder on this other host
oops i unplugged network
no idea.. I thought you meant "em1" as in the OpenBSD driver em(4)
nah this is ubuntu
Googling it... just, wtf, makes no sense.
Check dmesg to see what interfaces are enumerated and what they're renamed to, I guess.
on Debian, we have the clasic eth0, eth1, ..
there is a config file that keeps track of MAC address / interface name relationships
and reassigns them at startup
 /etc/udev/rules.d/70-net-persistent.rules or something like that
yea, that
seems like if I change the PCI slot of my Ethernet card, I still want the interface name to stay the same ..
http://pastebin.com/rEZB3EuR
it's weirder on the other host
i wonder what slot p18 is :)
Yeah that's my feeling. But I understand the other "descriptive" move - when dealing with faceless people in a datacenter far, far away, it's more important that they plug into the right physical port, and that if they replace a card, it keeps the same name.'
true
i always get confused when there are like 8 ethernet interfaces.
the biggest problem used to be is that different kernels would change what order they iterated through eth0, eth1, eth2 etc
so it wasn't static for onboard vs cards etc.
and so you upgrade kernels, and all the ordering changes.
And sometimes, depending on which bus enunerated first, they would move around on their own anyways.
openbsd's never had that problem for me.
Exactly
:D openbsd ftw
hmm i wonder if openbsd supports this network card...
What's the card?
mellanox connectx en
Hmmm
freebsd supports it from what i understand
(Yes it does)
I don't think OpenBSD supports infiniband at all
this is ethernet only
there's en, ib, and vpi models.  vpi do either en is only ethernet
Ah, good to know. Regardless, not seeing anything at all for "openbsd mellanox"
yeh i think it's not supported.
i wonder how much of the infiniband stuff it needs.
it could be a fun project to port freebsd driver to openbsd.
they're a lot cheaper than the intel cards.
althoguuh atm the only cheap ones are US shipping only :(
i've been trying to figure out if i can do 10gbe across my room