jokoon has quit [Read error: Connection reset by peer]
neurrowcat has joined #ipfs
<xelra>
Is my understanding correct that ipfs-cluster is an elaborate pin-bot?
infinitesum has quit [Quit: infinitesum]
<whyrusleeping>
xelra: its a bit more than that
<whyrusleeping>
ipfs-cluster has the all the same api calls as a regular ipfs node
<whyrusleeping>
but its smart about where each block gets put
cwahlers_ has quit [Read error: Connection reset by peer]
qjkxkcd has quit [Ping timeout: 245 seconds]
cwahlers has joined #ipfs
qjkxkcd has joined #ipfs
maxlath has quit [Ping timeout: 255 seconds]
[0__0] has joined #ipfs
<SchrodingersScat>
whyrusleeping: oh? so kinda like tahoe?
<whyrusleeping>
In some ways, yeah
<whyrusleeping>
though we don't yet have their awesome caps system
infinitesum has joined #ipfs
MDude has joined #ipfs
dconroy_ has joined #ipfs
jonnycrunch has quit [Ping timeout: 255 seconds]
jonnycrunch has joined #ipfs
infinitesum has quit [Quit: infinitesum]
[0__0] has quit [Remote host closed the connection]
bwerthmann has joined #ipfs
[0__0] has joined #ipfs
jkilpatr has quit [Ping timeout: 240 seconds]
droman has quit [Ping timeout: 260 seconds]
droman has joined #ipfs
infinitesum has joined #ipfs
MDude has quit [Ping timeout: 240 seconds]
Encrypt has joined #ipfs
jkilpatr has joined #ipfs
maxlath has joined #ipfs
infinitesum has quit [Quit: infinitesum]
parazyd has left #ipfs ["nuked."]
infinitesum has joined #ipfs
kants has quit [Ping timeout: 260 seconds]
kants has joined #ipfs
kants has joined #ipfs
kants has quit [Changing host]
gmoro_ has joined #ipfs
gmoro has quit [Ping timeout: 255 seconds]
Boomerang has quit [Quit: Lost terminal]
gmoro_ has quit [Read error: Connection reset by peer]
Mitar has quit [Ping timeout: 260 seconds]
Guest86515 has joined #ipfs
Guest46308 has quit [Ping timeout: 260 seconds]
espadrine has quit [Ping timeout: 240 seconds]
Mitar has joined #ipfs
citizenErased has joined #ipfs
maxlath has quit [Ping timeout: 245 seconds]
espadrine_ has joined #ipfs
Encrypt has quit [Quit: Quit]
<xelra>
What does the caps system do?
<lemmi>
apparently i'm in a google bubble. what's tahoe?
<detran>
tahoe-lafs
<xelra>
lemmi: I haven't used it yet, but tahoe-lafs is a system to combine storage.
<xelra>
In trustless systems.
<lemmi>
k
<SchrodingersScat>
it's pretty neat. I used to run a cluster of 4 nodes, used it as a pastebin
robattila256 has joined #ipfs
<xelra>
I've been procrastinating this for way too long, but I think today I'm gonna set up ipfs-cluster in combination with the private network feature.
<xelra>
See how well this performs as a replacment for traditional distributed fs.
<whyrusleeping>
xelra: please file issues on ipfs-cluster for anything wrong, broken or hard to figure out
wak-work has quit [Remote host closed the connection]
<whyrusleeping>
it works well enough when i tried it last, but its got a few sharp edges
<yuvipanda>
my IPFS chart for kubernetes is almost merged! I'll write a blog post on easy deployment of IPFS on kubernetes once that is done :)
<xelra>
Will do. Anything I need to be aware of when also running on some Windows nodes?
<xelra>
From experience, there's always a caveat to Windows. :)
wak-work has joined #ipfs
<whyrusleeping>
oooh, i have no idea
<whyrusleeping>
never tried it on windows
<whyrusleeping>
yuvipanda: great!
<SchrodingersScat>
caveat: Don't use windows.
infinitesum has quit [Quit: infinitesum]
<Atrus[m]>
Not even to escape a fire?
<Calinou>
use the best: Linux for servers, Mac for graphics, Windows for Solitaire
ianopolous has joined #ipfs
infinitesum has joined #ipfs
citizenErased has quit [Quit: Leaving]
<whyrusleeping>
there really isnt a good solitare program on linux
<whyrusleeping>
lets be honest here
<whyrusleeping>
the gnu one kinda sucks
<Xe>
as of win 10, not on windows either
<whyrusleeping>
whoa what
<Xe>
yeah
<Xe>
they fucked up solitare too :/
<whyrusleeping>
i'm so disappointed
<achin>
this is why you keep a desk of playing cards in your desk drawer :)
* whyrusleeping
doesnt have drawers in his desk :(
<achin>
that should be illegal
maxlath has joined #ipfs
<crankylinuxuser1>
yuvipanda: have you considered apache mesos? Ive been playing around with that ecosystem. Certainly interesting, for sure.
MDude has joined #ipfs
<whyrusleeping>
anyone feel like switching to blake2b and raw-leaves by default in 0.4.11?
sirdancealot has joined #ipfs
sirdancealot has quit [Read error: Connection reset by peer]
infinitesum has quit [Quit: infinitesum]
maxlath has quit [Ping timeout: 245 seconds]
bwerthmann has quit [Ping timeout: 272 seconds]
maxlath has joined #ipfs
<crankylinuxuser1>
whyrusleeping: you just read the HN article as well?
<whyrusleeping>
yeap
<whyrusleeping>
lol
<crankylinuxuser1>
i read that and lobste.rs
<crankylinuxuser1>
and about a dozen other sites I skim..
<whyrusleeping>
i havent gone on lobste.rs in so long
<whyrusleeping>
also: 'ipfs add --hash=blake2b-256 THING' works
<whyrusleeping>
in case anyone wants to try that
<whyrusleeping>
ipfs hashing a 910MB file normally takes 3.2 seconds
<whyrusleeping>
with blake2b, it takes 1.6
<whyrusleeping>
it appears to be exactly twice as fast
<ianopolous>
nice
<crankylinuxuser1>
wow, well damn. awesome :)
Encrypt has joined #ipfs
<whyrusleeping>
also, with raw leaves you get a lot of nice benefits
<whyrusleeping>
adding gets a little less cpu intensive, as you don't have to deal with protobuf in the leaves
<whyrusleeping>
(less memory intensive too)
<whyrusleeping>
and the blocks stored on disk are just raw pieces of the file
<whyrusleeping>
so it saves a small amount of diskspace
<yuvipanda>
crankylinuxuser1: yeah, we did an evaluation of mesos+marathon vs kubernetes a year and something ago
<yuvipanda>
kubernetes came out ahead so I ended up focusing on that heavily :)
intern has quit [Ping timeout: 260 seconds]
<whyrusleeping>
yuvipanda: we're working on a kubernetes ipfs thing to make testing ipfs in networked scenarios easier
<crankylinuxuser1>
good point. Things are just moving so quickly in the container and DCOS fronts. The only thing I keep seeing is "stay away from openstack"
<whyrusleeping>
if youre interested, help is always appreciated
<yuvipanda>
crankylinuxuser1: yeah, totally. you can run kubernetes on top of Mesos too
<yuvipanda>
haha, yes, re: openstack. I helped run one at $LASTJOB, never again
<whyrusleeping>
yuvipanda: oh wow, no, i hadnt seen that
<yuvipanda>
whyrusleeping: yeah, once that gets merged you can install ipfs on kubernetes with
CaptainPlanet has joined #ipfs
<yuvipanda>
helm install stable/kubernetes
<crankylinuxuser1>
heh, thats what we run, right now. kerberos/shibboleth/openldap/openradius runs on it....
<yuvipanda>
if you want 5 replicas you can do 'helm install stable/ipfs --set replicaCount=5'
<yuvipanda>
haha, I meant stable/ipfs not stable/kubernetes
<yuvipanda>
whyrusleeping: I'll add a patch with an ingress soon, so you can easily make a gateway for arbitrary domains that way too
sirdancealot has joined #ipfs
dimitarvp` has joined #ipfs
dimitarvp has quit [Disconnected by services]
dimitarvp` is now known as dimitarvp
ianopolous has quit [Ping timeout: 260 seconds]
CaptainPlanet has quit [Ping timeout: 260 seconds]
maxlath has quit [Ping timeout: 245 seconds]
maxlath has joined #ipfs
infinitesum has joined #ipfs
jkilpatr has quit [Ping timeout: 260 seconds]
<fluff[m]>
Can anyone point me in the right direction to solving the SSL error going on at https://ipfs.flufftronix.com ? I've got the DNS setup going and files are being served, just not without bypassing a browser warning first
<fluff[m]>
achin: yes I'm aware of that, I'm not sure how to resolve. the SSL is carrying over from the IPFS gateway
<achin>
i see. there is some special magic on ipfs.io that will let it "host" domains with a dnslink
<achin>
sounds like a question for someone like Kubuxu
* Kubuxu
was summoned
* achin
has magic Kubuxu summoning powers
<achin>
TIL
<Kubuxu>
there is no magic
<Kubuxu>
currently you can't do that with HTTPs
<Kubuxu>
we want to do that (by dynamically requesting and validating certs from LetsEncrypt)
<Kubuxu>
but not much is WIP right now
<achin>
have the gateways always been able to read the Host header to find the dnslink record?
<whyrusleeping>
yeap
<whyrusleeping>
for quite a long while
<achin>
neat, i didn't know that
<whyrusleeping>
yeah, its a pretty fun hack
jkilpatr has joined #ipfs
<achin>
so it's still magic, but just old magic :)
<whyrusleeping>
allowing users to piggyback on the main ipfs.io cert has weird security considerations
<whyrusleeping>
with ipfs.io/ipfs/SOMETHING its clear that we're just serving as a gateway
<whyrusleeping>
but with my.ipfs.domain
<whyrusleeping>
its less apparent
<achin>
indeed
<fluff[m]>
OK; lemme see if I can just get this going via http then. I know it's gone in and out of working without a browser warning as I've been tinkering
<infinitesum>
Is there any documentation on using ipfs.io as a gateway? I'm assuming that it's not meant to host production sites or whatever, just not sure if it's just meant to let people play with IPFS or if it can be used for more stuff
<whyrusleeping>
infinitesum: you can use it for your personal website and other small projects
<whyrusleeping>
but if youre going to do something commercial, or your plan on putting a ton of traffic on it, please talk to us
<whyrusleeping>
in that case you probably want to run your own gateway
<achin>
i bet there's no SLA on ipfs.io either (which is anothe rreason to run your own gateway)
<whyrusleeping>
^
<whyrusleeping>
yeap
<infinitesum>
@whyrusleeping ok that makes sense.
<whyrusleeping>
though our uptime is pretty great
<whyrusleeping>
i think we managed to get two nines last month
<achin>
well, you roundrobin ipfs.io, which is gonna help a lot
<Kubuxu>
we are also working on new infra
<whyrusleeping>
^
<Kubuxu>
under new domain
<whyrusleeping>
its gonna be awesome
<Kubuxu>
that will be anycast network
<fluff[m]>
OK, so http[s] problems avoided. Cool (and thanks!) On to the last issue I'm having: getting some clusters running so I don't have to manually re-pin every time I update
<whyrusleeping>
10gbps in multiple regions around the world
<achin>
hotness
<achin>
protocollabs seems to have an impressive bandwidth and storage budget
<whyrusleeping>
kevina: one other thing i just remembered, in namesys/republisher/repub.go, it grabs the private key from the peerstore, we should change it so it grabs that from the keystore
<whyrusleeping>
kevina: testing, let me look at a couple things
<kevina>
ok
<whyrusleeping>
kevina: on startup, the ipns entries will be republished after the republisher interval
<whyrusleeping>
which is set from the config
skeuomorf has joined #ipfs
callumgare1 has quit [Remote host closed the connection]
sprint-helper has quit [Write error: Broken pipe]
callumgare has joined #ipfs
sprint-helper has joined #ipfs
detran` has joined #ipfs
Mateon3 has joined #ipfs
Mateon1 has quit [Remote host closed the connection]
Mateon3 is now known as Mateon1
zuck05 has joined #ipfs
<kevina>
whyrusleeping: basically the goal is to have ipns keys generated with "ipfs key gen" behave the same as if the key was the node's key? Correct
<whyrusleeping>
exactly
<kevina>
okay that I can work with, there seam to be several problems that I will see if I can fix
<whyrusleeping>
sweet, thanks!
detran has quit [Ping timeout: 260 seconds]
Guest86515 has quit [Ping timeout: 260 seconds]
Guest86515 has joined #ipfs
<fluff[m]>
..so with ipfs-cluster, if I'm trying to pin a remote VPS to my local computer, which multiaddr that comes up via 'ipfs-cluster-ctl id' should I be using? I've tried a few and keep getting dial attempt fail errors
<kevina>
whyrusleeping: you said you want namesys/republisher/repub.go use the keystore. Should I try and change the key format from peer.ID to what the keystore uses?
<kevina>
it seams to me it will make everything easier
<whyrusleeping>
kevina: hrm... maybe we can add a 'GetByID' method to the peerstore?
<whyrusleeping>
er
<whyrusleeping>
keystore
<kevina>
so is there a good reason to keep using peer.iD?
<whyrusleeping>
a peer.ID is a pubkeyhash
<whyrusleeping>
and right now, the keystore just references keys by their human readable names
<kevina>
okay
<whyrusleeping>
peer.ID is just the canonical way of referencing a key
<kevina>
oh okay
<whyrusleeping>
probably should be something like key.ID instead, but its close enough
<kevina>
so how to a get a private key out of a peer.ID
<kevina>
without using the peerstore
<kevina>
looking at republishEntries
<whyrusleeping>
well, annoyingly enough, you have to check each key in the keystore to see if it matches right now
<whyrusleeping>
which is why i'm thinking we should add a helper method to the keystore that does that
rendar has quit [Quit: std::lower_bound + std::less_equal *works* with a vector without duplicates!]
<kumavis>
skylarnorris part of ipfs team?
<kumavis>
richardlitt: do you know
mahloun has joined #ipfs
<whyrusleeping>
kumavis: yeah
<whyrusleeping>
kumavis: i JUST had a question for you
<whyrusleeping>
any idea why geth would just randomly stop syncing?
<kumavis>
cool - does she have an irc handle?
<whyrusleeping>
uhm... its skylarnorris when shes on irc
<whyrusleeping>
but shes not normally on irc, need to get in touch with her?
<kumavis>
yeah she asked for some personal deets in github and want to DM them
<whyrusleeping>
sent you a PM
<kumavis>
ty
<kumavis>
whyrusleeping: why geth would stop syncing - idk, possibly ran out of some resource?
<whyrusleeping>
hrm... this keeps happening. I should probably post an issue on github
owlet has joined #ipfs
<kumavis>
we have a whole team to maintain our eth nodes so im unsure of the quirks
<kumavis>
asked them for tips tho
<whyrusleeping>
thanks <3
<fluff[m]>
Is there a way I can I resolve an IPNS address to an IPFS address on a machine that's not publishing the IPNS entry?
<whyrusleeping>
meanwhile, my bitcoin and zcash nodes havent needed to be touched in weeks
<fluff[m]>
(from terminal)
<whyrusleeping>
yeah
<whyrusleeping>
ipfs resolve /ipns/foobar
<whyrusleeping>
er
<whyrusleeping>
ipfs name resolve
<fluff[m]>
whyrusleeping: I'm getting: "Error: Could not resolve name."
<fluff[m]>
oh ok let me try that second one
<whyrusleeping>
fluff[m]: do you have a daemon running?
<whyrusleeping>
ipfs name resolve /ipns/foobar
<kumavis>
whyrusleeping: maybe you would prefer Parity?
<fluff[m]>
meh; same error.
<whyrusleeping>
i tried parity, it was actually worse
<kumavis>
> simondlr: There was a recent issue with it
<kumavis>
¯\_(ツ)_/¯
<whyrusleeping>
slower syncing, and (perhaps imagined) random api quirks
<fluff[m]>
whyrusleeping: yeah, daemon is running. I can run this command on my local machine and it works but on a remote one it won't resolve the locally hosted ipns
<whyrusleeping>
that resolves to /ipfs/QmaekS8fnJsPDVHJpjVN3KYfv8VQCxf99pdkUXaBFc2o9A for me
<whyrusleeping>
fluff[m]: what version of ipfs are you using?
<fluff[m]>
whyrusleeping: go-ipfs/0.4.9/
<whyrusleeping>
alright
<whyrusleeping>
and your node that isnt able to resolve that entry
<whyrusleeping>
how many peers does it have?
infinitesum has quit [Quit: infinitesum]
infinitesum has joined #ipfs
<fluff[m]>
not sure how to check on that
<fluff[m]>
actually looks like this must be an issue with the instances running via /etc/rc.local, when I kill those and restart the daemon I can resolve
pat36 has quit []
<fluff[m]>
OK both remote machines resolving correctly now
Foxcool has quit [Ping timeout: 246 seconds]
maxlath has quit [Quit: maxlath]
Caterpillar has quit [Quit: You were not made to live as brutes, but to follow virtue and knowledge.]
<fluff[m]>
Just added "ipfs pin add -r "$(ipfs resolve /ipns/QmV2GeY7Kvk9YxWt2YR5oK3kKtQm1wWmPcz7ccNYsJiaZv)" to my hourly cron .. no need to mess with clusters or manual re-pinning now.. right? If this does what I think it does (workaround to pin IPNS) this should be easily accessible info
owlet has quit [Ping timeout: 268 seconds]
owlet has joined #ipfs
infinitesum has quit [Quit: infinitesum]
appa has quit [Ping timeout: 260 seconds]
ethereumnerd has joined #ipfs
jaboja has joined #ipfs
ethereumnerd has quit [Client Quit]
Encrypt has quit [Quit: Quit]
mahloun has quit [Ping timeout: 240 seconds]
espadrine_ has quit [Ping timeout: 240 seconds]
<voker57>
if you don't mind old pins being left in place
<voker57>
and trust owner of this IPNS not to change link to 5 TB of data
ianopolous has joined #ipfs
<kevina>
whyrusleeping: in order for the repub to get the key from the keystore the "self" key needs to be stored in the keystore,
<kevina>
the question is where is the best place to do this?
<whyrusleeping>
hrm...
<whyrusleeping>
kevina: maybe for now, we can have a special case for "self" ?
<whyrusleeping>
its currently stored in the config
<whyrusleeping>
and we need to move it to the keystore
<kevina>
where to put the special case, there is already one in core.go
<kevina>
we need to test is the public key is the one for "self"
<kevina>
also it appears PutRecordToRouting is failing on the non "self" key. I am getting failed to republish: failed to find any peer in table
<kevina>
whyrusleeping: ^
<whyrusleeping>
maybe in the republisher, for now, we can say 'if id == OURPEERID { priv = node.OurPrivateKey }'
<whyrusleeping>
or something along those lines
<whyrusleeping>
hrm... failed to find any peer in table for just the other key? "self" works fine?
<kevina>
okay, cool, we need to store the "self" key, another special case
<kevina>
"failed to find any peer in table" I think "self" works fine, put am not certain yet
jkilpatr has quit [Ping timeout: 240 seconds]
achin has quit [Quit: WeeChat 1.7.1]
achin has joined #ipfs
jkilpatr has joined #ipfs
sirdancealot has quit [Ping timeout: 240 seconds]
dconroy_ has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<fluff[m]>
voker57: yes, that would just be for a trusted IPNS. I'm looking to host a site off my home computer, but also pin it to a couple of my VPS's for CDN purposes