<Ape>
Is anybody using ipfs as a backup utility? I'm thinking of combining ipfs with something like zbackup, then just pinning my data on multiple servers.
<mildred>
and I'm going to sleep soon. It's 1 am in France
<Ape>
zbackup for encryption and pre-encryption deduplication
<Ape>
Is there a way to verify that my data is being pinned on a number of servers without necessarily trusting those servers?
<Ape>
E.g. by making random queries to check that a random piece of my data is available
NightRa has quit [Quit: Connection closed for inactivity]
<jbenet>
mildred: will see that too. thank you so much for all the help here, and again sorry for not having a clearer resolution sooner
<Not_>
>Is there a way to verify that my data is being pinned on a number of servers without necessarily trusting those servers?
<Not_>
this would be sweet
<Not_>
knowing how many pins it had
<Not_>
like seeds
<mildred>
jbenet: thank you for looking up these PR.
<Not_>
Ape, you can open a github request
<achin>
if a gateway returns "context deadline exceeded" does it keep the hash in its wantlist?
<ion>
jbenet: (I'm sure you're sick of hearing variants of this joke.) Q: How many IPFS developers does it take to change a lightbulb? A: Just Juan.
anticore has quit [Ping timeout: 260 seconds]
jager_ is now known as jager
<Ape>
Not_: I will
<mildred>
I'll leave you to get some sleep. Good night :)
<Ape>
Also, does anybody have experience in using ipfs for this kind of personal backup? Is there any tools already available to make it easier?
Soft has quit [Read error: Connection reset by peer]
<feross>
whyrusleeping: you should join #arcticjs !
mildred has quit [Ping timeout: 250 seconds]
ygrek_ has quit [Ping timeout: 240 seconds]
ygrek has joined #ipfs
longguang has quit [Ping timeout: 245 seconds]
Gaboose has joined #ipfs
longguang has joined #ipfs
uhhyeahbret has quit [Quit: WeeChat 1.3]
uhhyeahbret has joined #ipfs
joshbuddy has quit [Quit: joshbuddy]
Soft has joined #ipfs
pfraze has joined #ipfs
reit has joined #ipfs
pfraze has quit [Remote host closed the connection]
joshbuddy has joined #ipfs
rendar has quit [Quit: std::lower_bound + std::less_equal *works* with a vector without duplicates!]
joshbuddy has quit [Quit: joshbuddy]
voxelot has quit [Ping timeout: 240 seconds]
rektide has quit [Ping timeout: 246 seconds]
warptangent_ has quit [Ping timeout: 240 seconds]
devbug has quit [Ping timeout: 245 seconds]
rektide has joined #ipfs
warptangent has joined #ipfs
<jakoby>
Ape: afaik nothing will pin unless you tell it to
<lgierth>
ipfs add pins
<jakoby>
so, we don't have the ability to pin on servers we don't have access to
<achin>
correct
<lgierth>
ipfs add pins
<lgierth>
oops
<jakoby>
could one just periodically load the data via the gateway to have the gateway hold on to the cache
<lgierth>
yeah
<jakoby>
cool
hashcore has quit [Ping timeout: 240 seconds]
<jakoby>
so, is it a known thing that the fuse mounts are slow and cpu intensive currently?
<achin>
there is no promises that the gateway will hold on to the cache for any period of time, though
<jakoby>
or should I submit an issue? :)
<achin>
best to submit an issue
Kalon has joined #ipfs
joshbuddy has joined #ipfs
pfraze has joined #ipfs
disgusting_wall has quit [Quit: Connection closed for inactivity]
Pria has quit [Ping timeout: 272 seconds]
Kalon has quit [Quit: Kalon]
autoxeny has joined #ipfs
<Ape>
jakoby: I can request somebody to pin my data. Then I want to verify if they did so. I can't make anybody to pin my data, but perhaps they are friendly or I could pay them.
<M-davidar>
ehd: if you're not careful it'll remove a limb
<ehd>
really, it should come with safety instructinos
<ehd>
*instructions
<patagonicus>
Calling ipfs files stat for every file in this mirror is … kinda slow. I think I'll have to switch to ipfs files ls + parsing for "production" use.
ygrek has joined #ipfs
<ipfsbot>
[go-ipfs] jbenet force-pushed v0.4.0-dev from 63a8e75 to 1c1f9c6: http://git.io/vRLhG
ygrek has quit [Ping timeout: 246 seconds]
pfraze has joined #ipfs
Senji has joined #ipfs
hoony has joined #ipfs
ygrek has joined #ipfs
hoony1 has joined #ipfs
chriscool has quit [Quit: Leaving.]
hoony has quit [Ping timeout: 255 seconds]
hoony1 is now known as hoony
chriscool has joined #ipfs
sivachandran has joined #ipfs
hartor has joined #ipfs
chriscool has quit [Read error: Connection reset by peer]
<dignifiedquire>
whyrusleeping: also quite a few binaries fail to build
lempa has quit [Ping timeout: 250 seconds]
<patagonicus>
dignifiedquire: Don't think I have the space right now to do so, was just gonna grab one or two videos.
<dignifiedquire>
whyrusleeping: also ipfs-update and fs-repo-migrations don't build anything for me
<dignifiedquire>
patagonicus: even two videos help :)
<patagonicus>
Hmm. All the HD ones seem to be 80G (just MP4 though, no WebM). Maybe I can find some space on one of my servers for that, but I want to get my Alpine mirror up first. rsync'ing to /ipns/local is very slow, unfortunately.
<daviddias>
seems it is processing some very large modules (150Mb ++)
<patagonicus>
Just noticed that cdn.media.ccc.de has an index which lists all files (including timestamp + size). All of their media seems to be just a bit more than 3TB. :D
<daviddias>
lgierth have you rebooted the node or something? Almost feels it is kind in a deadlock state for some reason (it was working just fine today's morning)
cSmith_ has quit [Changing host]
cSmith_ has joined #ipfs
cSmith_ is now known as cSmith
<dignifiedquire>
whyrusleeping: any thoughts on those logs?
<lgierth>
daviddias: didn't touch it
<lgierth>
it looks like it's still doing *something* though
<daviddias>
still has some space available too, must have locked for some other reason then
<lgierth>
well, no api requests though for the past 40 minutes
<daviddias>
pinged whyrusleeping to take a peek at it
<daviddias>
registry-mirror waits for ipfs to finish the write before trying to add more
<lgierth>
and a rise in goroutines just a minute before it stopped doing api stuff
<dignifiedquire>
lgierth: how is biham holding up? pinning is still running
<daviddias>
dignifiedquire: pinning of what?
<lgierth>
dignifiedquire: no idea at the moment :)
<dignifiedquire>
180gb stackexchange archives
<daviddias>
when did that pinning started?
<dignifiedquire>
2 days ago, but it was stopped over night as my machine was shutdown, somaybe 28h or so in total now
<dignifiedquire>
but it's veeery slow ipfs is uploading super slow from my machine :/ not even 10% of the capacity of my connection
pskosinski_ has joined #ipfs
pskosinski has quit [Read error: Connection reset by peer]
pfraze has joined #ipfs
chriscool has quit [Quit: Leaving.]
pfraze has joined #ipfs
zorglub27 has quit [Ping timeout: 245 seconds]
<daviddias>
so, that means that registry-mirror on biham is not only a couple of orders of magnitude faster
<daviddias>
but it was even fetching and verifying your stack exchange archive too?
<daviddias>
sweet! :D
<daviddias>
whyrusleeping check that ^^, o/ yeah
zorglub27 has joined #ipfs
<lgierth>
patagonicus: i'm going to mirror that to biham
<daviddias>
lgierth: biham seems to be locked now
<daviddias>
do you know what is the issue?
<lgierth>
no
<lgierth>
but i'm gonna fetch that goroutine dump for why
yosafbridge` has quit [Ping timeout: 250 seconds]
<patagonicus>
lgierth: Are you going to add that to IPFS yourself or will you just pin it once someone has done so?
chriscool has joined #ipfs
<lgierth>
patagonicus: gonna mirror and add it to ipfs
<patagonicus>
Nice.
<lgierth>
it doesn't make much sense fetch to 3 TB on a normal dsl connection haha
<lgierth>
or what have you got over there?
libman1 has joined #ipfs
libman has quit [Read error: Connection reset by peer]
gingeropolous has quit [Quit: Leaving]
<patagonicus>
Got a server with 100MBit and my dormitory got fibre to the uni which is our uplink. Not sure how fast it can be as I'm sharing with a lot of other people, but I've had ~60MByte/s down in the middle of a night once. :D
<patagonicus>
Upload is less of a problem as it's not used as much, but my problem is mostly that of storage. I could use some old drives, but that would mean losing some backups I have on there.
<whyrusleeping>
lgierth: don't reboot biham without telling daviddias first
<The_8472>
60MB/s on a 100Mbit link? impressive
<The_8472>
ah, silly me, dormitory and server being separate things
yosafbridge has joined #ipfs
<patagonicus>
Yeah. It's very easy to max out the servers bandwidth making it unresponsive for anything else. :D
<dignifiedquire>
daviddias: lgierth biham is still downloading refs for me
<lgierth>
that's a mirror run by a university in the same city as biham (nuremberg)
<lgierth>
not even sure if the bottleneck is network or disk io lol
anticore has joined #ipfs
<patagonicus>
lol
<Kubuxu>
lgierth: I my case (60MiB/s) it is the HDD, if I /dev/null/ data I get upto 95MiB/s :P
<Kubuxu>
HDD + cache
<Kubuxu>
it dropped then when cache filled
<lgierth>
yeah you're nearby too eh
<lgierth>
.pl?
<lgierth>
ok me -> supermarket
<lgierth>
whyrusleeping: i'm putting the dumps in /root
chriscool has quit [Ping timeout: 246 seconds]
<Kubuxu>
Yup.
<Kubuxu>
From my friend's business internet. I don't have that good one in my flat :/.
hartor1 has joined #ipfs
fsl has joined #ipfs
hartor has quit [Ping timeout: 245 seconds]
hartor1 is now known as hartor
The_8472 has quit [Ping timeout: 240 seconds]
jhulten has joined #ipfs
<dignifiedquire>
lgierth: niiiice
<dignifiedquire>
biham is getting pretty used these days :D
fsl has quit [Quit: leaving]
jhulten has quit [Ping timeout: 246 seconds]
<ScoreUnder>
I've got another query about IPFS 0.4. It seems like it leaks memory (it's currently using 1600MB, when it started around 400MB. I'd restarted it after a crash, which I can only assume must have been the OOM killer). Is this usage normal, and if not is there a way I can trace it back to the allocation like with valgrind or something? (I don't know Go enough to determine this kind of thing)
<ScoreUnder>
curently up to 1750M
ygrek has joined #ipfs
libman1 has quit [Ping timeout: 260 seconds]
anon3459 has quit [Ping timeout: 264 seconds]
ilyaigpetrov has quit [Quit: Connection closed for inactivity]
NightRa has joined #ipfs
joshbuddy has joined #ipfs
The_8472 has joined #ipfs
zorglub27 has quit [Ping timeout: 260 seconds]
libman has joined #ipfs
<lgierth>
ScoreUnder: if it runs out of memory on its own, it'll print a stacktrace
Soft- has joined #ipfs
<ScoreUnder>
lgierth: even taking into account linux's overcommit and such?
<ScoreUnder>
[1] 28003 killed ipfs -D daemon
<ScoreUnder>
ipfs -D daemon 26966.50s user 4755.62s system 53% cpu 16:31:09.19 total
<ScoreUnder>
↑ how it died
prosody is now known as prsdy
<ScoreUnder>
it's currently at 2158MB
<lgierth>
yeah i've seen it eat lots of ram too
<lgierth>
you might be seeing high numbers of goroutines too
<lgierth>
check :5001/debug/pprof
<lgierth>
the "full goroutine dump" or so is useful
<ScoreUnder>
it looks like 22118 goroutines are inside readHandshake ther
pfraze has quit [Remote host closed the connection]
warner has joined #ipfs
<dignifiedquire>
whyrusleeping: ping re distributions
joshbuddy has quit [Ping timeout: 256 seconds]
cemerick has quit [Ping timeout: 250 seconds]
joshbuddy has joined #ipfs
anticore has quit [Quit: bye]
jhulten has joined #ipfs
jaboja has joined #ipfs
cemerick has joined #ipfs
joshbuddy has quit [Quit: joshbuddy]
<Guest55618>
you could amke your own physical mesh network to repalace the internet. what hardwar/software would you use?
Soft- is now known as Soft
<The_8472>
physical as in wired?
<The_8472>
or physical as in not-internet-overlay?
<Guest55618>
physically wired
Guest55618 is now known as VegemiteToast
VegemiteToast is now known as VegemiteToast_
<VegemiteToast_>
not running through the internet.
<The_8472>
i'd get an excavator and then dig up all the gold needed to pay for the wiring
<VegemiteToast_>
I heard it was bad to daisy chain switches.
<VegemiteToast_>
and then?
<VegemiteToast_>
It's hypothetical money is not an obstical
<VegemiteToast_>
if every only pays for half the price to connect to their neighbor/peer it probably wouldn't cost that much
<VegemiteToast_>
long distance it would
<ScoreUnder>
VegemiteToast_: and I suppose human nature is not an obstacle either? I doubt it would ever happen because it would require lots of people locally to buy into it, and that requires everyone being a techie. Others will be happy with their client-server subscriber model internet access and I'm sure ISPs would do all they can to keep the mesh separate
<The_8472>
digging fiber trenches to each of your neighbors + some box to route potentially gigabits of traffic is not that cheap
<Kubuxu>
VegemiteToast_: dark fibers and FPGAs.
<VegemiteToast_>
u really thing every peer wouldd need fiber?
<The_8472>
and even if money is not an issue, you still need human labor. you could hire a whole nation to do the digging and wiring...
<VegemiteToast_>
think*
<The_8472>
well, who wants slow internet?
<Kubuxu>
Not every but many, WiFI links have high latency.
<VegemiteToast_>
1 gigabit ethernet. with each peer server atleast to others.
<VegemiteToast_>
serving*
<The_8472>
well, he said "physically wired", which rules out wifi anyway
<VegemiteToast_>
two*
<The_8472>
everyone also need to pipe other people's data through their nodes, not just their own
<Kubuxu>
And almost everywhere where new infrastructure fiber links are laid.
<VegemiteToast_>
ye.
<The_8472>
since it's a mesh it'll flow through many many hops
<VegemiteToast_>
yip
<The_8472>
and you would need some serious optical switching otherwise the optical-electronic conversion on hundreds and hundreds of hops will kill latency
<VegemiteToast_>
but caching data on nodes would reduce hopage.
<The_8472>
then you also need storage
<ScoreUnder>
VegemiteToast_: you're thinking of a web replacement rather than an internet replacement then
<VegemiteToast_>
would kill voip and live shit though
<VegemiteToast_>
ye. I suppose
<The_8472>
just think how a path from china to south america would look like
<VegemiteToast_>
so u would have to have an exchange in every suburb just like ISPs?
<VegemiteToast_>
to reduce hops
<ScoreUnder>
then you could connect all the nearby exchanges together into one super-exchange... then again, and you've reinvented the ISP :^)
<Kubuxu>
If you were to create global mesh, separate from internet, you still need long range high capacity links.
<The_8472>
or the non-profit IXP at least
<The_8472>
we have a bunch of those here in europe
<The_8472>
but of course they can only accept large customers, i.e. ISPs, individuals would overwhelm them
<The_8472>
i guess you could kinda have IXPs-in-small in some way... but that still requires digging lots of trenches
<Kubuxu>
ScoreUnder: no, if you choose different networking design, cjdns for example.
<VegemiteToast_>
are cjdns hops any faster?
<The_8472>
Kubuxu, a mesh is nice locally. but how to you connect countries and continents?
<The_8472>
or even cities for that matter
<VegemiteToast_>
international and national fiber backbone is a must
<The_8472>
meshes pretty much end at the city boundary
<VegemiteToast_>
but locat mesh is probably do-able
<VegemiteToast_>
local*
<VegemiteToast_>
question is can u have a mix of the two. and how would u do it?
<The_8472>
but given unlimited resources we could just move all humans into a single city!
<Kubuxu>
The_8472: you use long range fibers and open market.
<The_8472>
Kubuxu, you'd still need someone to build and run them. at best you can replace for-profit ISPs with non-profit infrastructure providers
<VegemiteToast_>
If one city made a meshnet they could sell access to their mesh to the next city over.
<Kubuxu>
Partially the same way current internet is laid out as on global scale internet looks like mesh.
<VegemiteToast_>
it could be profitable
<The_8472>
VegemiteToast_, nobody wants to do the accounting
<The_8472>
accountants waste human life-hours and money
<VegemiteToast_>
hehe
<Kubuxu>
The_8472: noone says about non profit, cjdns will be including pay per packet in undefined future (when we figure it out).
<VegemiteToast_>
or make installation of a nodes hardware, quid pro quo, If you server files for me. I'lll install it for free.
<Kubuxu>
VegemiteToast_: Berlin has really nice local mesh if you are interested.
<The_8472>
that's close to how peering and "I put my CDN cache into your server rack" deals work
libman has quit [Ping timeout: 260 seconds]
<VegemiteToast_>
is it viable?
<The_8472>
Kubuxu, in which case would paying for packets have some utility?
<VegemiteToast_>
after the initial investment of wardware. the customer would serve files for the rest of his/her life.
<VegemiteToast_>
in excahnge for access to the mesh.
<Kubuxu>
you need money even if not for profit then for upgrades and maintenance.
<Kubuxu>
also: non-profit != free
<The_8472>
i know
joshbuddy has joined #ipfs
<VegemiteToast_>
make people pay you to become a certified tech dude
<The_8472>
but you can do that by paying for the hardware, not for the routing
<VegemiteToast_>
then u have workers
mildred has joined #ipfs
<VegemiteToast_>
hi i'm john and I'm meshnet certified worker
<VegemiteToast_>
badge no. 1337
<VegemiteToast_>
make them get recertified every year.
<VegemiteToast_>
lol
<VegemiteToast_>
sound like a pyramid scheme but its not
<Kubuxu>
What if I move around the world and I just want to use network? Also PPP has nice benefit of limiting attack and stimulating infrastructure development.
<VegemiteToast_>
ppp?
<Kubuxu>
Pay per Packet
<ScoreUnder>
(I was wondering what on earth point to point protocol had to do with it)
<Kubuxu>
sorry for that
<VegemiteToast_>
if u moved around the world u could be certified to set up your own node. or get some one to come out and set one up. or get payed to convince others to be a node.
<The_8472>
Kubuxu, well, idk, paying for capacity sounds nicer psychologically. you just pay a flat fee per day/month/year
<The_8472>
if you pay for traffic instead of capacity you worry about your usage
<VegemiteToast_>
just make it.. if u serve more traffic ur price goes down.
<Kubuxu>
The_8472: it is hard to pay per capacity as you don't know whom to pay
<The_8472>
but solving the easy problem causes psychological anguish
<Kubuxu>
in pay per packet you just pay ones you route through.
<Kubuxu>
It isn't easy problem, it is impossible problem, now in the internet you have backbone to which the capacity is calculated, in mesh you don't have that.
<Kubuxu>
also everything depends what is the price.
<VegemiteToast_>
what I really want is a way to server files to the guy next door. but still be able to connect accross the world with reasonalbe ping
<ScoreUnder>
VegemiteToast_: "Oy bill hop on my wifi a sec and I'll send it over"
<VegemiteToast_>
without having to interact with people
jaboja has quit [Ping timeout: 250 seconds]
<VegemiteToast_>
and changing networks all the time
<VegemiteToast_>
I want a dual network situation
<Kubuxu>
And that is what global mesh is about, you have planty of local links then you have hubbing and few intercity, -coutry, -continental.
<VegemiteToast_>
is cjdns really up to the job atm?
<Kubuxu>
VegemiteToast_: cjdns already offers that. You have global network but you can easily setup local mesh that will be part of that global network.
<VegemiteToast_>
and u need a cjdns open-wrt router?
<VegemiteToast_>
routers*
<Kubuxu>
no you can use your PC, notebook anything that runs anything
<Kubuxu>
for a mesh local you probably need a dedicated router (possibly RouterBoard) with antenna and so on.
<VegemiteToast_>
how do u pysically wire to another cjdns computer and to the internet?
<The_8472>
why do you want to physically wire something?
<The_8472>
wlan is easier on short distances
<VegemiteToast_>
speed
Encrypt has quit [Quit: Quitte]
<The_8472>
directed antenna?
<VegemiteToast_>
are they significantly faster?
<The_8472>
so you can use <newest standard> between two houses
<Kubuxu>
You need someone that already has connection to Hyperboria to connect to the rest Hyperboria.
<Kubuxu>
only problem of WiFi is about 2-10ms of latency.
<VegemiteToast_>
what like the best sorta range u can get out of the latest wireless routers?
<VegemiteToast_>
is 50 meters doable?
<Kubuxu>
With dedicated antenna?
<VegemiteToast_>
dedicated.. as in directional?
<VegemiteToast_>
i suppose ye
<Kubuxu>
as big external.
zorglub27 has joined #ipfs
<VegemiteToast_>
oh. ye. what are they like?
<VegemiteToast_>
is it viable?
<Kubuxu>
With clear line of sight up to 2-5km
<VegemiteToast_>
huh. and the other end needs a big ass antenna too?
<The_8472>
VegemiteToast_> oh. ye. what are they like? <- whatever tradeoff between size and distance you want to make
<VegemiteToast_>
huh. and the other end needs a big ass antenna too?
<The_8472>
if you want to connect two houses next to each other you obviously don't need a dish
libman has joined #ipfs
<VegemiteToast_>
isn't it illeage to transmit that far without permission
<Kubuxu>
They are about 70cm 20cm 20cm
<Kubuxu>
no
<The_8472>
depends on your jurisdiction and the power involved
<Kubuxu>
range and size depends on type of antenna
hellertime has joined #ipfs
cemerick has quit [Ping timeout: 264 seconds]
chriscool has joined #ipfs
rombou has joined #ipfs
patcon has joined #ipfs
rombou has quit [Ping timeout: 264 seconds]
pfraze has joined #ipfs
chriscool has quit [Quit: Leaving.]
patcon has quit [Ping timeout: 256 seconds]
chriscool has joined #ipfs
simonv3 has joined #ipfs
NightRa has quit [Quit: Connection closed for inactivity]
joshbuddy has quit [Quit: joshbuddy]
zorglub27 has quit [Read error: Connection reset by peer]
kahiru has joined #ipfs
zorglub27 has joined #ipfs
anticore has joined #ipfs
chriscool has quit [Ping timeout: 265 seconds]
pfraze has quit [Remote host closed the connection]
joshbuddy has joined #ipfs
<whyrusleeping>
dignifiedquire: sup
<dignifiedquire>
whyrusleeping: still waiting for a fix to build migrations and update
<whyrusleeping>
bork bork bork
<whyrusleeping>
dignifiedquire: uhm... not sure why it failed
<whyrusleeping>
what does 'make ipfs-update' yield?
<whyrusleeping>
dignifiedquire: does dists/ipfs-update/tag_versions have anything in it/
<whyrusleeping>
?
libman has quit [Remote host closed the connection]
<whyrusleeping>
after i finish listening to this song
disgusting_wall has joined #ipfs
<ScoreUnder>
Since I'm sure you'd be thrilled to see how nasty it looks, pure sh version: remove_lines() { file1=$1 file2=$2; set --; while IFS= read -r line; do set -- "$@" "$line"; done <"$file2"; while IFS= read -r line; do found=false; for line_b do [ "$line_b" = "$line" ] && found=true && break; done; $found || printf %s\\n "$line"; done <"$file1"; }; remove_lines A B
<whyrusleeping>
tperson: doesnt work on OSX for some reason
<tperson>
Works on my laptop :D
<whyrusleeping>
ScoreUnder: that looks like tasty marmalade
<ipfsbot>
[js-ipfs] diasdavid pushed 3 new commits to feat/bootstrap: http://git.io/vuFm7
<ipfsbot>
js-ipfs/feat/bootstrap 292e75f David Dias: refresh repo for each battery of tests
<ipfsbot>
js-ipfs/feat/bootstrap f404a8f David Dias: mend
<ipfsbot>
js-ipfs/feat/bootstrap af42da7 David Dias: mend
<ScoreUnder>
I hope it is. Bought the value and non-value for comparison. Because my life is that interesting :D
<whyrusleeping>
tperson: odd...
<whyrusleeping>
even if B is empty?
<tperson>
Nope, not that lol
<whyrusleeping>
haha
<whyrusleeping>
i've purchased more licorice than i'd care to admit to
<whyrusleeping>
to compare
<whyrusleeping>
i hate how much latency i've got right now...
<dignifiedquire>
whyrusleeping: alright I got a folder full of stuff for ipfs-update, need the same fixes for fs-repo-migrations I guess
<whyrusleeping>
they should already be made
* whyrusleeping
is going to sleep
<tperson>
OSX should fix there shit, that is really annoying that grep fails if the file supplied by -f is empty
<jgraef>
I'm creating a lot of DAG nodes and my network is clogged. So I guess that outbound traffic comes from inserting keys into the DHT?
<The_8472>
is the bandwidth saturated or are you seeing dropped packets?
<The_8472>
packet drops could indicate conntrack table saturation. turning off conntrack or reducing timeouts in the nat device could help. assuming you have control over it
pfraze has joined #ipfs
<jgraef>
As far as I can tell the bandwidth is not saturated. I'm not behind a NAT (I'm in the university's VPN).
<The_8472>
well, still could be a stateful firewall
robcat has joined #ipfs
robcat has quit [Client Quit]
<dignifiedquire>
whyrusleeping: okay cool :)
janosch2 has joined #ipfs
janosch2 is now known as jgraef2
<jgraef2>
I disconnected the VPN and now everything is fine again :)
jgraef has quit [Ping timeout: 256 seconds]
zorglub27 has quit [Quit: zorglub27]
jgraef2 is now known as jgraef
<dignifiedquire>
whyrusleeping: got another bug :P dist.json is empty for fs-repo-migrations
mildred has quit [Ping timeout: 256 seconds]
hartor1 has joined #ipfs
hartor has quit [Ping timeout: 264 seconds]
hartor1 is now known as hartor
jgraef has quit [Ping timeout: 245 seconds]
<achin>
if a gateway returns "context deadline exceeded" does it keep the hash in its wantlist?