stebalien changed the topic of #ipfs to: Heads Up: To talk, you need to register your nick! Announcements: go-ipfs 0.4.19 and js-ipfs 0.34 are out! Get them from dist.ipfs.io and npm respectively! | Also: #libp2p #ipfs-cluster #filecoin #ipfs-dev | IPFS, the InterPlanetary FileSystem: https://github.com/ipfs/ipfs | Logs: https://view.matrix.org/room/!yhqiEdqNjyPbxtUjzm:matrix.org/ | Forums: https://discuss.ipfs.io | Code of
litvintech has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
vmx has quit [Remote host closed the connection]
reit has quit [Ping timeout: 250 seconds]
The_8472 has quit [Ping timeout: 252 seconds]
Elon_Satoshi is now known as Copenhagen_Bram
The_8472 has joined #ipfs
<cyberwolf[m][m]>
whether to do the gateway?
placer14 has joined #ipfs
Pulse2496 has joined #ipfs
i1nfusion has quit [Remote host closed the connection]
pie_ has joined #ipfs
<pie_>
i dont know much about ipfs yet but people seem to regard it highly
<pie_>
how are merge conflicts handled?
<c0c00n[m]>
if you're talking about content, it is referenced by its unique hash, its immutable, so no conflict to take care of... if you're talking of other things like pubsub etc..., IPFS makes use of CRDTs (quick example: 1+2+3=5 as 2+3+1=5) which is conflict-free too at the end
mowcat has joined #ipfs
<pie_>
(im superficially familiar with CRDTs as a concept)
<pie_>
hm ok, it kind of sounded to me like it might need to handle merges or something, which sounded sketchy. I guess I'll need to look into IPFS more sometime and find out whats actually going on :)
i1nfusion has joined #ipfs
Pulse2496 has quit [Ping timeout: 246 seconds]
i1nfusion1 has joined #ipfs
xnaas has quit [Quit: Dead.]
xnaas has joined #ipfs
<c0c00n[m]>
depends of what you think it can do :)
i1nfusion1 has quit [Remote host closed the connection]
<pie_>
no idea honestly but i would have been surprised if it had some totally magical merge resolution :p
cris_thor` has joined #ipfs
gmoro_ has joined #ipfs
cris has quit [Ping timeout: 246 seconds]
BenLubar has quit [Ping timeout: 255 seconds]
BenLubar has joined #ipfs
spinza has quit [Ping timeout: 255 seconds]
kivutar_ has joined #ipfs
matt-h has quit [Ping timeout: 246 seconds]
Ll1i1lL has quit [Ping timeout: 246 seconds]
gmoro has quit [Ping timeout: 246 seconds]
blallo has quit [Ping timeout: 246 seconds]
kivutar has quit [Ping timeout: 255 seconds]
matt-h has joined #ipfs
blallo has joined #ipfs
Ll1i1lL has joined #ipfs
spinza has joined #ipfs
vmx has joined #ipfs
<m_tadeu>
so pin works like a refcount? (still trying to understand how all this works)
<m_tadeu>
even if I pin a file, that file doesn't exist in my node?
<c0c00n[m]>
pinning mainly means that the corresponding content (blocks) to the pinned hash won't get deleted on garbage collection
<c0c00n[m]>
so I suppose that it means you already have a local copy of that content
<c0c00n[m]>
but haven't checked
<m_tadeu>
so ipfs files ls /ipfs/Qm... should show me the dir contents (Qm... being the cid to a dir that is pinned)
chiui has quit [Ping timeout: 272 seconds]
chiui has joined #ipfs
<c0c00n[m]>
To simplify (a lot), on one side when you add content to IPFS, your local node just split the file/data into blocks and keep a representation of it through a Merkle tree (tree of hashes, blocks being leaf), which means that you can guaranty blocks/files/data integrity checking those hashes. On the other side, each node rely on all the libp2p network stack in order to get in touch. So when you request a content (by its hash),
<c0c00n[m]>
each node which has that content will start sending corresponding blocks that compose that content, that's why if a file/data/content is hosted on many nodes, it's fetched quite fast as there's a lot of download sources.
<c0c00n[m]>
On most file systems directories are inodes, which means it's like a file which content is a list of contained entries which can be files or other inodes/subdirectories. So filenames are just a directory's information. And directories are data too so you can have of hash of it too.
<c0c00n[m]>
That's why IPFS can claim it deduplicate contents as those are referenced by their hash and not only a directory reference. For example, if you create two files both with the same 'hello world' content, under your current OS file system you'll have two referenced file file1.txt and file2.txt in the same directory, but with IPFS you'll have a directory hosting references 'file1.txt' and 'file2.txt' both pointing to the same
<c0c00n[m]>
content hash as the content was the same.
<c0c00n[m]>
Now if you remember that contents are splitted into blocks, theorically, it should improve ressources consumption to store data as no doublons will be stored.
dqx_ has joined #ipfs
<m_tadeu>
yeah I get the hasing->content concept...it's pretty cool :)
<m_tadeu>
I'm experimenting an ipfs private network as a tinny cluster with my machines as nodes
<m_tadeu>
pinning worked great, so now I have several directories pinned on all machines
<m_tadeu>
but I was expecting 'ipfs files ls /' to show the pinned directories
<m_tadeu>
ipfs files == mutable file system, is this correct?
Soo_Slow has joined #ipfs
lnostdal has quit [Excess Flood]
ylp has quit [Quit: Leaving.]
lnostdal has joined #ipfs
<c0c00n[m]>
yes
<c0c00n[m]>
ipfs files ls [<path>] - List directories in the local mutable namespace.
<c0c00n[m]>
can you give a try to 'ipfs file ls' command ?
<c0c00n[m]>
ipfs file ls /ipfs/
<m_tadeu>
Error: path must contain at least one component
<c0c00n[m]>
ipfs pin ls
<c0c00n[m]>
?
<m_tadeu>
ahh hashes showd up
<m_tadeu>
but on another node won't return anything
<m_tadeu>
dispite it has the same pins
<c0c00n[m]>
pinning commands were executed on other nodes ?
<m_tadeu>
well I'm using the cluster for that
<m_tadeu>
but I see them in ipfs pin ls
<c0c00n[m]>
looks like it's not auto pinning, maybe check your config?
<m_tadeu>
what is auto pinning? where is that configured?
<m_tadeu>
anyway, it's pinning on all nodes....just mfs won't see them
<Voker57>
mfs is not related to pinning
vitz3 has joined #ipfs
<m_tadeu>
only to added files, then?
manray has quit [Ping timeout: 250 seconds]
dirkmc has quit [Excess Flood]
dirkmc has joined #ipfs
i1nfusion1 has joined #ipfs
plexigras has joined #ipfs
clemo has quit [Ping timeout: 255 seconds]
clemo has joined #ipfs
<Voker57>
added files are automatically pinned, yes
<Voker57>
but still are not related to MFS
<Voker57>
MFS is, like, a convenient directory editor
<Voker57>
and stuff in MFS is also protected from GC
<c0c00n[m]>
let me guess: MFS is somewhat your local file system, without a copy of your file into your local IPFS node repository. Am I right?
manray has joined #ipfs
lnostdal has quit [Ping timeout: 250 seconds]
lnostdal has joined #ipfs
detran has joined #ipfs
renich has joined #ipfs
<Voker57>
no
<Voker57>
MFS is virtual file system, index over your IPFS repo
<Voker57>
one that you create yourself
The_8472 has quit [Ping timeout: 240 seconds]
<m_tadeu>
and files managed by MFS won't be pinned but won't be removed by GC either...?
<Voker57>
yes
vmx has quit [Remote host closed the connection]
The_8472 has joined #ipfs
chiui has quit [Ping timeout: 255 seconds]
afx has quit [Quit: No Ping reply in 180 seconds.]
nighty- has quit [Quit: Disappears in a puff of smoke]
eyenx has quit [Ping timeout: 246 seconds]
^andrea^1492 has joined #ipfs
^andrea^149 has quit [Ping timeout: 246 seconds]
Soo_Slow has quit [Remote host closed the connection]
Soo_Slow has joined #ipfs
pie_ has quit [Ping timeout: 245 seconds]
__jrjsmrtn__ has joined #ipfs
rendar has quit []
_whitelogger has joined #ipfs
Jesin has quit [Quit: Leaving]
Jesin has joined #ipfs
vitz3 has quit [Remote host closed the connection]
vitz3 has joined #ipfs
i1nfusion1 has quit [Remote host closed the connection]
i1nfusion1 has joined #ipfs
tombusby has quit [Remote host closed the connection]
tombusby has joined #ipfs
The_8472 has quit [Ping timeout: 252 seconds]
<Voker57>
don't think so but public ipfs gateways performance is poor in my experience
<Voker57>
and on cloudflare one videos are blocked altogether
The_8472 has joined #ipfs
pie_ has joined #ipfs
<Mikaela>
How poor would you consider it compared to small upload bandwidth or OnionShare?
<Voker57>
dunno, it's more about not being able to connect to some peers and stalling for long time, bw when it actually downloads can be up to 1 mbs
<Mikaela>
Thanks, I hope the recepient will get the video and I am not made to learn video editing/compressing/something to learn how can it be 11 GB and how to make it smaller :)
<Mikaela>
(I was also let understand that it was videoed by a not-high-end-smartphone which makes me more confused about the size)
<rialtate[m]>
Probably not very compressed at all
<rialtate[m]>
Is it possible to set up a relay that only serves pinned links?
<Mikaela>
I think it's possible, but am not sure on the config