<okket>
maybe it's easier to work on a udp based protocol than try to fix/optimize the current tcp based?
<xelra>
luzifer: Pushover. It's the best since Notify.
<whyrusleeping>
okket: i'm just working on simulating it so i can get reliable performance numbers of the system
<okket>
tcp is good for a small number of reliable connections where order of packets are paramount. is that really the case for ipfs? mostly sending metadata around sounds to me more like job for UDP, like DNS
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<whyrusleeping>
okket: we do want to use udp
<whyrusleeping>
but we need reliable streams
<whyrusleeping>
we dont want any dropped packets
<ipfsbot>
[go-ipfs] whyrusleeping pushed 1 new commit to feat/netsim: http://git.io/vJ8vC
<ipfsbot>
go-ipfs/feat/netsim a2d55ae Jeromy: WIP: a slightly different approach to the network simulator
Daboloskov has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
Daboloskov has quit [Client Quit]
therealplato1 has quit [Quit: Leaving.]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
andrewtj has quit [Quit: Connection closed for inactivity]
notduncansmith has joined #ipfs
AtnNn has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
dgomez10 has joined #ipfs
<dgomez10>
hello
Daboloskov has joined #ipfs
Daboloskov has quit [Client Quit]
therealplato has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
hellertime has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
Daboloskov has joined #ipfs
Daboloskov has quit [Client Quit]
www has quit [Ping timeout: 256 seconds]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
tilgovi has quit [Remote host closed the connection]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
dgomez10 has quit [Ping timeout: 246 seconds]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
Tv` has quit [Quit: Connection closed for inactivity]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
mrkeon1 has joined #ipfs
patcon has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
mrkeon1 has quit [Quit: WeeChat 1.1.1]
sharky has quit [Ping timeout: 264 seconds]
hellertime has quit [Quit: Leaving.]
EricJ2190 has quit [Ping timeout: 246 seconds]
Daboloskov has joined #ipfs
sharky has joined #ipfs
Daboloskov has quit [Client Quit]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<whyrusleeping>
quiet day...
patcon has quit [Ping timeout: 244 seconds]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<whyrusleeping>
okket: i think i have an idea about those context switches
philholden has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<ipfsbot>
[go-ipfs] whyrusleeping created fix/ticker-waste (+1 new commit): http://git.io/vJ8DR
<ipfsbot>
go-ipfs/fix/ticker-waste 867186c Jeromy: dont create a new ticker each loop
notduncansmith has joined #ipfs
inconshreveable has quit [Remote host closed the connection]
notduncansmith has quit [Read error: Connection reset by peer]
<whyrusleeping>
jbenet: ping
<ipfsbot>
[go-ipfs] whyrusleeping opened pull request #1180: dont create a new ticker each loop (master...fix/ticker-waste) http://git.io/vJ8yf
inconshreveable has joined #ipfs
<ipfsbot>
[go-ipfs] whyrusleeping created fix/mega-read (+1 new commit): http://git.io/vJ8yF
<ipfsbot>
go-ipfs/fix/mega-read 517e2d1 Jeromy: quick fix for OOM panic that has been plaguing us
<ipfsbot>
[go-ipfs] whyrusleeping opened pull request #1181: quick fix for OOM panic that has been plaguing us (master...fix/mega-read) http://git.io/vJ8yp
<whyrusleeping>
cool, lets hope we dont see any more of those issues :)
<whyrusleeping>
i want my ipfs nodes to stay up longer than a few days!
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<whyrusleeping>
okket: do you know much about travisCI?
<whyrusleeping>
i'm trying to figure out if it runs our tests in parallel on the same machine
<okket>
nope, only looked as logs to find out why things did not work
<whyrusleeping>
hrm, thats about where i'm at too
<whyrusleeping>
lol
Daboloskov has joined #ipfs
Daboloskov has quit [Client Quit]
philholden has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
www has joined #ipfs
Marklock has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
philholden has quit [Ping timeout: 250 seconds]
grncdr has joined #ipfs
philholden has joined #ipfs
grncdr1 has joined #ipfs
grncdr has quit [Ping timeout: 272 seconds]
philholden has quit [Ping timeout: 246 seconds]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
ir2ivps9 has quit [Ping timeout: 250 seconds]
Daboloskov has joined #ipfs
Daboloskov has quit [Client Quit]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<crest>
is `ipfs daemon` expected to use 100% of one core on a FX-8350 to write ca. 500kB/s to the /ipns fuse mountpoint?
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<whyrusleeping>
crest: thats not something i've observed
<whyrusleeping>
but how big of a file are you adding?
<crest>
whyrusleeping: about 700MiB
<whyrusleeping>
hrm, i'm not super surprised its taking that much CPU
<whyrusleeping>
its not what i expect, but the codepath for adds isnt super optimized yet
<crest>
any tips on where to start profiling?
<whyrusleeping>
the code that youre invoking by writing to /ipns is in unixfs/mod
<crest>
or is it possible to add files with ipfs and link them into my /ipfs/<node-id>?
<whyrusleeping>
essentially, your data is being passed through the kernel, to fuse, to our fuse code, and then to the dagmodifier
<crest>
*ipfs add
<whyrusleeping>
crest: of course!
<whyrusleeping>
i'm actually working right now on commands to make that easier
<whyrusleeping>
but you can ipfs add a file, then 'ipfs name publish <hash>'
<whyrusleeping>
and it will show up as your ipns entry
<whyrusleeping>
do you also mind filing an issue about that operation being slow and eating CPU?
<crest>
whyrusleeping: thx i looked for something like ipfs ln oder flags to ipfs add
<whyrusleeping>
anything that doesnt feel right, or isnt outright userfriendly and apparent is a documentation bug IMO, so feel free to file bugs as you see fit :)
<crest>
whyrusleeping: as soon as i know more than "it feels too slow"
<whyrusleeping>
cool, glad to have someone poking around in that code
<crest>
and i just gave ipfs a second try now that the golang fuse bindings support freebsd
<whyrusleeping>
ooOooo, someone using freeBSD! woo!
<whyrusleeping>
that is uncharted territory for us
<crest>
i was too busy/lazy to properly fix the fuse bindings. i just ripped the multiplatform support out and defined the right types last time
<crest>
now everything builds properly on a clean freebsd 10.1/amd64 with go 1.4.2 from ports
<whyrusleeping>
sweet!
<whyrusleeping>
good to hear
<whyrusleeping>
anyways, its 3:30am here, i'm going to get some sleep
<crest>
and while busy with writes the /ipns/<node-id> fuse mountpoint takes multiple seconds to list a directory with 10 entries
<crest>
but it works
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
Daboloskov has joined #ipfs
notduncansmith has joined #ipfs
Daboloskov has quit [Client Quit]
notduncansmith has quit [Read error: Connection reset by peer]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
Marklock has quit [Ping timeout: 246 seconds]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
ei-slackbot-ipfs has quit [Remote host closed the connection]
ei-slackbot-ipfs has joined #ipfs
nickemery has joined #ipfs
nickemery is now known as ehmry
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
Daboloskov has joined #ipfs
philholden has joined #ipfs
Daboloskov has quit [Client Quit]
grncdr1 has quit [Quit: Leaving.]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
EricJ2190 has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
mildred has quit [Read error: Connection reset by peer]
mildred has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
mildred has quit [Read error: Connection reset by peer]
Marklock has joined #ipfs
inconshreveable has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
Marklock has quit [Ping timeout: 246 seconds]
Marklock has joined #ipfs
mildred has joined #ipfs
Daboloskov has joined #ipfs
Daboloskov has quit [Client Quit]
Marklock has quit [Ping timeout: 246 seconds]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<ipfsbot>
[go-ipfs] wking opened pull request #1183: core/commands/internal/slice_util: Remove this unused package (master...remove-core-commands-internal-slice_util) http://git.io/vJBbZ
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<ipfsbot>
[go-ipfs] wking opened pull request #1184: core/corenet: Remove this unused package (master...remove-core-corenet) http://git.io/vJReN
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<ipfsbot>
[go-ipfs] wking opened pull request #1185: Add package comments to packages that are missing them (master...go-doc) http://git.io/vJRUF
Marklock has joined #ipfs
RzR has quit [Excess Flood]
RzR has joined #ipfs
Marklock has quit [Ping timeout: 246 seconds]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
Daboloskov has joined #ipfs
Daboloskov has quit [Client Quit]
jck has quit [Ping timeout: 256 seconds]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
philholden has quit [Ping timeout: 240 seconds]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
Evermore has quit [Ping timeout: 240 seconds]
Evermore has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
grawity has joined #ipfs
<crest>
if i add a file with ipfs add $path how can i add a link to it to /ipns/$nodeid without writing it through the fuse mountpoint again?
atrapado has joined #ipfs
<AtnNn>
crest: ipfs name publish
<crest>
AtnNn: that just publishes it under a random? name
<AtnNn>
no, it uses your node id as the name
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<crest>
AtnNn: it doesn't work for me
<crest>
i expect `ipfs name publish $hash` to publish a mapping from my node id to the hash
<crest>
but nothing changes
<crest>
instead i get a new name with no way to link to it from my root directory
anshukla has joined #ipfs
<zorun>
crest: it should work, but there is some kind of cache
Daboloskov has joined #ipfs
<whyrusleeping>
ipfs name publish does not work while the fuse mountpoint is active
<whyrusleeping>
ive been meaning to add an error message or something to the command to say this
Daboloskov has quit [Client Quit]
<grawity>
where could I insert a function call to run as soon as `ipfs daemon` has finished starting the API server?
<crest>
whyrusleeping: that explains my problems
notduncansmith has joined #ipfs
<jbenet>
crest: would you mind filing an issue?
<jbenet>
o/
notduncansmith has quit [Read error: Connection reset by peer]
therealplato has quit [Read error: Connection reset by peer]
therealplato has joined #ipfs
<whyrusleeping>
jbenet: welcome to the land of the living!
<jbenet>
\o
<jbenet>
am at a wedding. actually at the place that nodeconf usually happens.
<crest>
is it possible to patch the root e.g. insert this subtree instead of adding everything again and publishing the new root?
<crest>
lets say i have a folder with 200 rather long podcast episodes and want to add episode 201
<okket>
this would clash with my understanding how ipfs works. the root hash must change if anything below changes. this is the way how DAG work, at least in my limited understanding
<crest>
okket: thats how a DAG works
grncdr has joined #ipfs
<crest>
but the you can reuse subgraphs to build the new root
notduncansmith has joined #ipfs
<crest>
instead of "deduplicating after the fact"
notduncansmith has quit [Read error: Connection reset by peer]
<okket>
i'd build my model so that such non-changing subgraphs are generated by default, e.g. by month folders...
<okket>
and not put all episodes in one huge dir ;)
<crest>
i'm looking for something like zfs
<crest>
if some node in the dag changes
<crest>
all nodes along the path to the root are replaced by new nodes and a new root node is created
<crest>
ipfs feels more like creating a new filesystem and filling it from a "backup"
<okket>
a distributed network filesystem ;)
philholden has joined #ipfs
<okket>
as far as I understand IPFS at its core works as you describe, how this translates to IPNS and how this then gets distributed to other nodes, I don't know.
<jbenet>
crest: that's exactly how IPFS works
<jbenet>
crest: it's like git AND xfs
<jbenet>
err zfs*
<jbenet>
(too close on the kb)
<crest>
jbenet: so this just not implemented yet?
<jbenet>
crest: it is implemented, just the tooling is not fully there.
<jbenet>
you can craft an object manually with `ipfs object get/put`, and create nodes that way
<whyrusleeping>
crest: what you describe is actually what i was working on last night
notduncansmith has quit [Read error: Connection reset by peer]
ehmry has quit [Quit: Leaving]
nickemery has joined #ipfs
nickemery is now known as ehmry
<Bat`O>
is there a nice way to check if an ipfs daemon is running ?
<Bat`O>
ping an http api point ?
<jbenet>
Bat`O -- try /version on either the API or Gateway
<Bat`O>
k
Marklock has joined #ipfs
<ipfsbot>
[go-ipfs] whyrusleeping created feat/save (+1 new commit): http://git.io/vJ0Cv
<ipfsbot>
go-ipfs/feat/save f4a74ab Jeromy: initial implementation of save
* whyrusleeping
kachow!
<cryptix>
save?
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
Daboloskov has joined #ipfs
<Marklock>
cryptix: +1
<cryptix>
marklock: "Save a hash to a location within your ipns keyspace",
Daboloskov has quit [Client Quit]
<Marklock>
Good to know, thanks!
<cryptix>
whyrusleeping: how is this different from publish?
grncdr has quit [Quit: Leaving.]
ehmry has quit [Quit: Leaving]
<cryptix>
god the cmdlib is really hairy..
* cryptix
tableflips and gets a beers
Marklock has quit [Ping timeout: 246 seconds]
<whyrusleeping>
cryptix you can specify a path within your keyspace
<whyrusleeping>
instead of overwriting it entirely
<whyrusleeping>
I'm not sold on the name
<whyrusleeping>
but the functionality is killer
grncdr has joined #ipfs
<ipfsbot>
[go-ipfs] jbenet closed pull request #1163: Record Datastore metrics for flatfs and leveldb (master...ds-metrics) http://git.io/vJe40
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<jbenet>
whyrusleeping: agreed we need that, but dont think it fits in the top level, nor the name. also-- it may be useful to model manipulation of objects using a staging tree, like git.
<whyrusleeping>
huh?
<whyrusleeping>
i know the names not right
<whyrusleeping>
i just needed to call it something so i could write it
<okket>
should 'ipfs publish' not figure out automatically if you only replace a subtree?
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<whyrusleeping>
okket: hrm, we could modify publish to do this functionality
<ipfsbot>
[go-ipfs] whyrusleeping opened pull request #1187: initial implementation of save (master...feat/save) http://git.io/vJ0V5
<jbenet>
cryptix: i know-- we should define interfaces and rewrite the important parts
notduncansmith has quit [Read error: Connection reset by peer]
inconshreveable has quit [Ping timeout: 256 seconds]
nickemery has joined #ipfs
nickemery is now known as ehmry
Daboloskov has joined #ipfs
Daboloskov has quit [Client Quit]
<whyrusleeping>
i really want to be able to sit at a coffee shop and have mdns pick up other peoples ipns records, and then present them to me so i can browse the 'local web'
anshukla has joined #ipfs
<whyrusleeping>
we can totally do this already, but just needs a bit more work on making it trivial
<crest>
whyrusleeping: also great for hackerspaces ;)
<whyrusleeping>
yes!
<whyrusleeping>
did anyone here play megaman (battle network) games?
<substack>
I'm not sure how I feel about that feature, seems to have a ton of privacy implications
<whyrusleeping>
that game had local internet and the internet of things before either were popular
<whyrusleeping>
substack: if you publish your ipns entry, its public anyways
<substack>
that's true
<substack>
but public on the internet is different from public in close local proximity
<whyrusleeping>
hrm?
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<whyrusleeping>
substack: how do you mean?
ehmry has quit [Quit: Leaving]
<substack>
you know the identities of people in close physical proximity and what they've published
RzR has joined #ipfs
<whyrusleeping>
if someone is worried about being discovered for one reason or another, they can simply disable mdns
<grawity>
on a related note, does ipfs' mdns code work when the system already has Avahi running? will they conflict?
<whyrusleeping>
grawity: i have not actually tested this
<whyrusleeping>
the mdns code is at the 'it works for what weve used it for' stage of testing
<grawity>
hmm, I'll test at work maybe
<whyrusleeping>
but not much goes wrong if it doesnt work, aside from maybe an error message or two
<grawity>
well, either way, it's better than flooding the network with broadcasts
<grawity>
which even new projects seem to keep doing :|
* whyrusleeping
is glad we dont do that
<whyrusleeping>
it actually works really nicely, at the hackerspace i hang out in here, ill frequently get connected to friends as soon as the turn their laptops on
inconshreveable has joined #ipfs
atrapado has quit [Quit: Leaving]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
mildred has quit [Quit: Leaving.]
mildred has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
fleeky has quit [Read error: Connection reset by peer]
fleeky has joined #ipfs
mildred has quit [Quit: Leaving.]
Daboloskov has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
lgierth has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]