jkilpatr_ has quit [Remote host closed the connection]
Jesin has quit [Ping timeout: 255 seconds]
heimindanger has quit [Quit: Page closed]
ccii has joined #ipfs
Jesin has joined #ipfs
Dawnbreaker[m] has joined #ipfs
deikiyoch has joined #ipfs
deikiyoch has left #ipfs [#ipfs]
Mateon1 has quit [Ping timeout: 240 seconds]
CrAzYPiLoT has quit [Quit: See you in hell.]
CrAzYPiLoT has joined #ipfs
CrAzYPiLoT has quit [Changing host]
CrAzYPiLoT has joined #ipfs
erikj- has quit [Ping timeout: 240 seconds]
erikj- has joined #ipfs
_whitelogger has joined #ipfs
pomegranatedaddy has joined #ipfs
<pomegranatedaddy>
hello
dPow has quit [Remote host closed the connection]
<pomegranatedaddy>
<pomegranatedaddy>
<pomegranatedaddy>
exit
pomegranatedaddy has quit [Client Quit]
pomegranatedaddy has joined #ipfs
espadrine has quit [Ping timeout: 240 seconds]
pomegranatedaddy has quit [Ping timeout: 240 seconds]
dPow has joined #ipfs
binarycat has joined #ipfs
chris6131 is now known as chris613
dhruvbaldawa has joined #ipfs
dhruvbaldawa has quit [Ping timeout: 240 seconds]
sim590 has joined #ipfs
<sim590>
When I launch the IPFS daemon with `ipfs daemon` (ipfs is the binary coming from the go-ipfs downloadable dist tarball), I get the error: ERROR dht: checking dht client type: context canceled notif
<sim590>
.go:81
<sim590>
is it normal from time to time? Am I missing something?
pomegranatedaddy has joined #ipfs
dylanPowers has joined #ipfs
dPow has quit [Ping timeout: 252 seconds]
knix has quit [Ping timeout: 252 seconds]
knix has joined #ipfs
knix has quit [Changing host]
knix has joined #ipfs
pomegranatedaddy has quit [Ping timeout: 240 seconds]
<Icefoz_>
Never seen that one.
cdata_ is now known as cdata
<binarycat>
This libp2p thing... It seems extremely abstract and difficult to grasp. I have no clue what any of these headers mean in the README.
<binarycat>
I "just" want to be able to send messages back and forth without using a server.
<Icefoz_>
And, in my (still limited) experience, a lot of IPFS is very much in the realm of "we're too busy changing how it works to write how it works"
Alpha64 has quit [Read error: Connection reset by peer]
pomegranatedaddy has quit [Ping timeout: 255 seconds]
upperdeck has quit [Ping timeout: 248 seconds]
upperdeck has joined #ipfs
ulrichard has joined #ipfs
Mateon2 has joined #ipfs
girrrrrrr2 has joined #ipfs
jonnycrunch has quit [Ping timeout: 248 seconds]
jonnycrunch1 is now known as jonnycrunch
jonnycru1ch has joined #ipfs
pawalls has quit [Remote host closed the connection]
kvda has joined #ipfs
pawalls has joined #ipfs
<binarycat>
Icefoz_: Very worrying.
<grawity>
are you expecting it to be set in stone by now
<grawity>
I'd sure like to see more documentation, but trying to fully specify a protocol before seeing if it's implementable rarely works...
<grawity>
so I'm just gonna wait
<Icefoz_>
I mean, basic IPFS functionality seems to generally work by now. Hence why they're working on hammering out specs.
<Icefoz_>
Which is fine, as far as it goes.
rendar has joined #ipfs
pomegranatedaddy has joined #ipfs
<Icefoz_>
I personally really want a lightweight IPFS server for a commercial project that does nothing but get, add and pin objects, and the existing go-ipfs program is *really not* that
NullConstant has quit [Read error: Connection reset by peer]
<Icefoz_>
So I'm torn between trying to write my own, and just diving in to help mop it up a bit.
}ls{ has quit [Quit: real life interrupt]
pomegranatedaddy has quit [Ping timeout: 240 seconds]
<kevina>
wrong channel, lets take this to ipfs-dev :)
erictapen has quit [Ping timeout: 255 seconds]
<daviddias>
ChrisMatthieu: sounds awesome :)
<daviddias>
ChrisMatthieu: what kind of presentation are you planning? Can it be in the next all hands?
erictapen has joined #ipfs
bwn has joined #ipfs
girrrrrrr2 has joined #ipfs
chris6131 has joined #ipfs
strauberry has quit [Read error: Connection reset by peer]
Encrypt has joined #ipfs
strauberry has joined #ipfs
pomegranatedaddy has joined #ipfs
knix has quit [Quit: leaving]
pomegranatedaddy has quit [Ping timeout: 240 seconds]
knix has joined #ipfs
knix has quit [Changing host]
knix has joined #ipfs
dhruvbaldawa has joined #ipfs
dhruvbaldawa has quit [Ping timeout: 248 seconds]
<ChrisMatthieu>
daviddias: Next all-hands on Monday, October 23rd would be perfect. I was just thinking about introducing myself, team, and computes.com. Explain how we are leveraging IPFS instead of the blockchain for decentralized and distributed computing. If that works, I could probably cover the intro in 10-15 minutes and take questions along the way.
redfish has joined #ipfs
<Alpha64>
the uber for computing quip is pretty bad
<daviddias>
ChrisMatthieu: sounds good :)
ianopolous_ has quit [Read error: Connection reset by peer]
<ChrisMatthieu>
daviddias: Is the call at 9am Pacific?
<daviddias>
ChrisMatthieu: the IPFS All Hands issue will be open tomorrow
<ChrisMatthieu>
Alpha64: I agree. Especially since we are targeting B2B - not B2C
<Alpha64>
it's just that "the uber for" phrase is so tired
<ChrisMatthieu>
so true
maxlath1 has joined #ipfs
maxlath has quit [Ping timeout: 255 seconds]
maxlath1 is now known as maxlath
ianopolous_ has joined #ipfs
neo[m]1 has joined #ipfs
atrapado_ has joined #ipfs
mentos1386 has joined #ipfs
Jesin has quit [Quit: Leaving]
kirby__ has joined #ipfs
pomegranatedaddy has joined #ipfs
muvlon has joined #ipfs
Jesin has joined #ipfs
kirby__ has quit [Ping timeout: 240 seconds]
pomegranatedaddy has quit [Ping timeout: 258 seconds]
kirby__ has joined #ipfs
<jamesstanley>
help! my ipfs gateway is being spammed with requests for /ipns/QmZPP1yxBVijWoAiNDSyXPwT2N8hiwzn1ReFx2Q8VnDTgT/cc.txt
<jamesstanley>
does anyone know what is going on?
<jamesstanley>
it's ipfs.jes.xxx aka hardbin.com
ianopolous_ has quit [Ping timeout: 255 seconds]
<jamesstanley>
there are tens of requests per second for that url, from lots of different ips - is someone ddos'ing my server?
<victorbjelkholm>
jamesstanley: do you have any proxy in between? Recommended way would be to have nginx or something in between, and block the remote that is requesting that.
<victorbjelkholm>
is just from one IP or many different ones?
<jamesstanley>
yes I have nginx in between
<jamesstanley>
it's from many ips, I can't spot any pattern
<lgierth>
hey
<victorbjelkholm>
hm, temporary solution would be to block "/ipns/QmZPP1yxBVijWoAiNDSyXPwT2N8hiwzn1ReFx2Q8VnDTgT"
<lgierth>
that's malware and likely part of a botnet command and control
<lgierth>
we blocked that on the ipfs.io gateway
<jamesstanley>
blocked by the url?
<jamesstanley>
i.e. /ipns/QmZPP1yxBVijWoAiNDSyXPwT2N8hiwzn1ReFx2Q8VnDTgT/* ?
<lgierth>
jamesstanley: let us know whenever that happens again, or other malware-ish stuff -- you can sign up with google webmaster tools and they'll notify you whenever google adds any of your URLs to the SafeBrowsing block list
whenisnever has joined #ipfs
<lgierth>
e.g. pretty much anything ending in .exe i've seen on ipfs.io so far was malware
<jamesstanley>
what might be more entertaining is to return executables that disable the botnet client
<lgierth>
hehe yeah
<lgierth>
we've refrained from that since we don't wanna break the content-addressing
<jamesstanley>
thanks for your denylist - I'll probably include it verbatim
<lgierth>
obliged :)
Encrypt has quit [Quit: Quit]
<jamesstanley>
do you think on balance it would be helpful or unhelpful if I were to do some more digging and then write a blog post about how botnets are using ipfs?
<jamesstanley>
it might have the unfortunate effect of making people associate ipfs with crime
<jamesstanley>
but apart from that I think it would be interesting
<lgierth>
it would just be really important to be careful
<lgierth>
it's very easy to take sentences out of context
<lgierth>
super interesting though, agreed -- i'll over review and advice :)
<fractex>
you also wouldn't want to write something like that without talking about what is being done to resolve the issue
<lgierth>
yep definitely
<lgierth>
we're working this on a couple of different fronts
<lgierth>
and the tl;dr for botnet operators is still that ipfs isn't a good fit because it'll expose part of their infrastructure
<lgierth>
i just think some haven't noticed yet :)
<jamesstanley>
lol
<lgierth>
actually don't take my word for it being botnet command-and-control -- who knows! i didn't examine it, just ran things through AV etc.
<jamesstanley>
it sounds totally plausible to me
<jamesstanley>
but I wouldn't write anything without doing more investigation
<fractex>
how would you go about investigating that comprehensively, anyway?
<jamesstanley>
I don't know yet, I'd have to wait until I got my hands on some of the content
<lgierth>
ipfs hasn't been optimized towards anonymity yet -- so there are clearnet addresses everywhere
<jamesstanley>
currently it's non-resolvable
<jamesstanley>
most of the requests are for cc.txt but there's the odd one for cloudnet.exe, under the same ipns name
<lgierth>
i might or might not have shot down the source of that hash you're looking for via an abuse report a month ago
<jamesstanley>
the cool thing about ipns is they can pop up somewhere else and as long as they have the key, they can start serving content again
<lgierth>
yeah! :)
<fractex>
how very cool
sim590 has quit [Ping timeout: 258 seconds]
gde33 has quit [Ping timeout: 255 seconds]
lassulus has quit [Quit: WeeChat 1.9.1]
sim590 has joined #ipfs
lassulus has joined #ipfs
dhruvbaldawa has joined #ipfs
AgenttiX has quit [Remote host closed the connection]
dhruvbaldawa has quit [Ping timeout: 240 seconds]
<Icefoz_>
The less cool thing is they're tied to a particular node instead of a logical name, so must have that key no matter what. And so there's no way to do clustering or indirection of your server (such as proxies or failover systems or load balancers), afaict. And keys are not human-readable.
<Icefoz_>
Actually, what DOES happen if you have two IPFS nodes with the same key and thus the same IPNS ID?
<Icefoz_>
If you ask to resolve that ID do you just get whichever one your own node happens to find first? Or does the network not let that happen?
<jamesstanley>
my understanding is that ipns links are published in the dht, so you'd just get whatever was published most recently
<jamesstanley>
but there's a high probability that I am wrong or missing something
shizy has quit [Ping timeout: 264 seconds]
<Icefoz_>
Huh, so it actually uses the pubsub mechanism?
<Icefoz_>
That's kinda cool
<jamesstanley>
no, just the dht
<jamesstanley>
but again, I'm liable to be talking nonsense :)
<Icefoz_>
I'm a bit grumpy about IPNS 'cause it doesn't actually solve any problems I am interested in, but that's still neat.
<jamesstanley>
what problem did you want it to solve?
<Icefoz_>
Pretty much what I said above.
<Icefoz_>
Those are problems that, as far as I can tell, IPFS has no realistic solution to.
<lgierth>
IPNS doesn't use pubsub yet, but soon
<lgierth>
right now it just writes signed pointers into the dht
<lgierth>
and these can be written by anyone with the matching private key
<Icefoz_>
It starts with "human readable names that can be updated effectively instantly" and everything kinda goes from there.
<jamesstanley>
Icefoz_: you want it to help you with "proxies, failover systems, load balancers"? If so, I think your mental model of ipfs is not accurate
<lgierth>
i think they mean failover when it comes to publishing that specific IPNS record
<lgierth>
you can have multiple nodes publishing to the same ipns key
<lgierth>
`ipfs key` and `ipfs name publish -k`
<Icefoz_>
Okay I need to understand this more clearly then.
<Icefoz_>
I publish an IPNS name on my one node. Then that node goes offline. Then someone else tries to look up that key. What happens?
<Icefoz_>
I guess I should try it and find out.
<lgierth>
the record will be valid for 12h or so
<lgierth>
someone with the matching private key needs to stay online republishing it
<lgierth>
i'm just saying it can be multiple nodes sharing that private key, so redundancy is given
<Icefoz_>
Hmmm, interesting. Okay that makes life somewhat easier.
<lgierth>
just don't use the node's actual identity for it
<lgierth>
use `ipfs key` instead
drathir has quit [Remote host closed the connection]
<Icefoz_>
What's the latency between a new value for the key being published and it being visible to random nodes?
<lgierth>
not very good yet? :D
<Icefoz_>
All right. :D
<lgierth>
that's why pubsub-based IPNS is in the works
<Icefoz_>
Well, that's still more useful than I thought it was, thank you.
<Icefoz_>
Speaking of useful, I want to use IPFS for an infrastructure project and have a bunch of things talking to it through the HTTP API, but after running for a couple hours the server seems to get very flaky (timeouts, etc). Where do I start contributing to make it more stable?
<lgierth>
find out more about what you're node is doing
<lgierth>
e.g. ipfs swarm peers
thujyl has joined #ipfs
<lgierth>
i suppose it's just way over 2000 peers when it gets slower
<lgierth>
we just merged connection management to mitigate that
<lgierth>
the network grew a ton recently :/
<lgierth>
there's also a patch in the works that significantly reduces CPU load
pomegranatedaddy has joined #ipfs
<lgierth>
and then the :5001/debug/pprof endpoints are useful too for getting insight
<lgierth>
these work with the standard pprof tools in the go toolchain
<lgierth>
we're using 1500/2000/1m here at the moment, but still experimenting
<Icefoz_>
Well, it is being flaky and it seems to be at 1995 peers or so, so. CPU reduction will be nice too. ;-) I'll try the latest master and see if that makes a difference.
<Icefoz_>
I'm not too hot at Go but I want IPFS to be production-quality in terms of robustness, at least when it comes to basic features, so if I can help I will be happy to.
pomegranatedaddy has quit [Ping timeout: 260 seconds]
<lgierth>
whyrusleeping might have some issues for starters
<Icefoz_>
If I can reliably pin and serve a couple terabytes of data to anything that asks for it I'll be happy.
<Icefoz_>
All right, thanks.
mingc has quit [Quit: Ping timeout (120 seconds)]
<lgierth>
about "a couple terabytes", try the badger datastore, it has much better performance with large datasets than the standard flatfs datastore
<Icefoz_>
Any interest in an Amazon block storage datastore? :D
mentos1386 has quit [Ping timeout: 240 seconds]
<lgierth>
there's an s3 datastore somewhere i think? :) veeeery slow though
drathir has joined #ipfs
<Icefoz_>
Heheh. Well, I'll poke around and experiment.
rodolf0 has quit [Ping timeout: 246 seconds]
mentos1386 has joined #ipfs
drathir has quit [Remote host closed the connection]
drathir has joined #ipfs
mentos1386 has quit [Client Quit]
mentos1386 has joined #ipfs
mentos1386 has quit [Remote host closed the connection]
kus_ubuntui686 has joined #ipfs
onikuma has quit [Quit: Leaving]
<victorbjelkholm>
think it's this one: https://github.com/ipfs/go-ds-s3 old and very slow unless you run daemon and your s3 bucket is in the same zone (still slow compared to local repo)
* victorbjelkholm
has not tried go-ds-s3 himself, so might not even build
<Icefoz_>
Hey, it has two whole commits, how bad could it be?
muvlon has quit [Quit: Leaving]
<Icefoz_>
Might be interesting to play with at least, if I can get it to build at all.
muvlon has joined #ipfs
rcat has joined #ipfs
jonnycrunch has joined #ipfs
atrapado_ has quit [Quit: Leaving]
bawitdaba has joined #ipfs
kirby__ has quit [Remote host closed the connection]
ccii has quit [Ping timeout: 255 seconds]
dummy has joined #ipfs
dummy is now known as Guest19251
jonnycrunch has quit [Ping timeout: 248 seconds]
Guest19251 has quit [Client Quit]
pomegranatedaddy has joined #ipfs
ccii has joined #ipfs
pomegranatedaddy has quit [Ping timeout: 252 seconds]
ccii has quit [Ping timeout: 255 seconds]
cxl000 has quit [Quit: Leaving]
stoopkid__ has quit []
stoopkid has joined #ipfs
<jdev>
does anyone know a good, decentralized db where the user doesnt have to store the whole db / blockchain?
ccii has joined #ipfs
<substack>
jdev: that depends a lot on the specifics of the use case you have in mind
<substack>
here are some questions: where does the database need to run? what kind of data will be stored? where does the data need to go?
<substack>
how partition-tolerant does the database need to be?
jdev has quit [Ping timeout: 258 seconds]
ZaZ has quit [Quit: Leaving]
bawitdaba has quit [Quit: WeeChat 1.9.1]
dhiaa has joined #ipfs
kus_ubuntui686 has quit [Ping timeout: 240 seconds]
dhiaa has left #ipfs [#ipfs]
noffle has left #ipfs ["WeeChat 1.9-dev"]
maxlath has quit [Quit: maxlath]
dhruvbaldawa has joined #ipfs
binarycat has joined #ipfs
ashark has quit [Ping timeout: 258 seconds]
ericst has quit [Remote host closed the connection]
dhruvbaldawa has quit [Ping timeout: 240 seconds]
Aranjedeath has joined #ipfs
thujyl has quit [K-Lined]
Aranjedeath has quit [Client Quit]
pomegranatedaddy has joined #ipfs
pomegranatedaddy has quit [Ping timeout: 240 seconds]
muvlon has quit [Quit: Leaving]
Aranjedeath has joined #ipfs
jaboja has joined #ipfs
erictapen has quit [Ping timeout: 240 seconds]
erictapen has joined #ipfs
erictapen has quit [Remote host closed the connection]