Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
    Peter Blockman
    @peterblockman

    Hello, I am building a react app working with orbit-db. I am having this error Error:undefinedis not supported by the IPLD Data Model and cannot be encoded
    I initialize IPFS like this https://github.com/ipfs-shipyard/ipfs-share-files/blob/2d3aecdbc694e41d3151e09526865eb5c88334cb/src/bundles/ipfs-provider.js#L44 with the latest version of ipfs-provider, ipfs-http-client and ipfs-core

    "ipfs-http-client": "^54.0.1",
     "ipfs-provider": "^2.1.0",
    "ipfs-core": "0.12.2"
     "orbit-db": "^0.28.0",
      "orbit-db-store": "^4.3.2",

    Any idea? thank you in advance

    pyc
    @xphorm:matrix.org
    [m]
    hello
    anyone tell me how to make a code block in this chat?
    1 reply
    PeacefulAF
    @PeacefulAF
    can I use web3 storage with orbitdb and how to connect it?
    pyc
    @xphorm:matrix.org
    [m]
    Actually someone told me you can use ```js
    And then in the new line you put code
    anders
    @tabcat00:matrix.org
    [m]
    @peterblockman: that usually happens when you are setting undefined or null to a value that is being encoded and uploaded to ipfs/ipld. you should check to make sure you arent doing that with whatever orbitdb store you are using
    Prasad Karmalkar
    @prasadkarmalkar
    Hello I am new to orbit db. I was trying to build sample application with orbit db but when i create app with node js and try to connect that database with its address using control center data which i had inserted from app is not displaying in control center
    someone please help me with this
    anders
    @tabcat00:matrix.org
    [m]
    you may need to add things to your ipfs config that enable the node to discover and connect to peers
    Prasad Karmalkar
    @prasadkarmalkar
    const ipfsOptions = {
    repo: './ipfs2',
    EXPERIMENTAL: {
    pubsub: true
    }
    }
    THIS IS MY IPFS CONFIG
    breadhunter
    @EmiM

    Hello, I have two issues:

    1) We just switched to new orbit-db and I noticed that the behaviour of replicated event changed.

    I have a simple test - one node adds 1000 entries to its local log db, second node connects and replicates this db.
    Previously (orbit-db < 0.27) I was receiving replicated event every ~32 entries (smaller chunks).
    Now I get only one big chunk and one replicated event after all 1000 entries are done being replicated.

    Is it an expected behaviour? Is there a way to send smaller chunks as before?

    2) What times of replication I should expect after upgrading orbit-db, ipfs etc.?
    Currently we experience ~20 seconds for localhost p2p network and ~150 seconds for tor network:
    https://github.com/ZbayApp/waggle/runs/4326485402?check_suite_focus=true (both 1000 entries).

    The times are very similar to what was before:
    https://github.com/ZbayApp/waggle/runs/3705573421?check_suite_focus=true

    mfidelman
    @mfidelman
    I just went through the "Field Guide" and for the life of me, I couldn't find anything that actually described the algorithms used for the different kinds of OrbitDB CRDTs. Can someone point me to a "theory of operations" document akin to the original IPFS paper, or something? Thanks!
    mfidelman
    @mfidelman:matrix.org
    [m]
    I just went through the "Field Guide" and for the life of me, I couldn't find anything that actually described the algorithms used for the different kinds of OrbitDB CRDTs. Can someone point me to a "theory of operations" document akin to the original IPFS paper, or something? Thanks!
    mfidelman
    @mfidelman
    Thanks!
    Holmes Wilson
    @holmesworcester:matrix.org
    [m]
    anders: any thoughts on emi's question above about the replicated event?
    Holmes Wilson
    @holmesworcester:matrix.org
    [m]
    I'm also curious if any work is being done to improve syncing performance, or if it's understood where the bottlenecks are.
    anders
    @tabcat00:matrix.org
    [m]
    there may be an option to keep them as smaller chunks, let me check. the changes made in 0.28 were made to improve sync performance
    Holmes Wilson
    @holmesworcester:matrix.org
    [m]
    Thanks! Also do you have any internal metrics on sync performance or some kind of test suite for this? It would be great to be able to compare our results to that. We haven't seen any improvement at least since the summer I think, in our own tests.
    @EmiM: do you have any historical performance data from our tests? Have the results changed at all across different OrbitDB versions?
    anders
    @tabcat00:matrix.org
    [m]
    Holmes Wilson
    @holmesworcester:matrix.org
    [m]
    What's the performance increase you've seen from these changes?
    anders
    @tabcat00:matrix.org
    [m]
    ill ping haad for emis question since i dont think the option controls that part anymore
    Holmes Wilson
    @holmesworcester:matrix.org
    [m]
    Thanks!
    anders
    @tabcat00:matrix.org
    [m]
    i havent ran them unfortunately
    1 reply
    Holmes Wilson
    @holmesworcester:matrix.org
    [m]
    I think the thing we are trying to achieve with the replicated event is to see new transactions as they are received instead of when sync is complete. Is that right @EmiM ?
    anders
    @tabcat00:matrix.org
    [m]
    yes
    breadhunter
    @EmiM
    yes, that's right. We send data to the frontend on each 'replicated' event (e.g send messages). I am not sure if user wants to wait for all 1000+ messages to replicate before he sees anything.
    I would want to understand how it changed since orbit-db 0.26
    Holmes Wilson
    @holmesworcester:matrix.org
    [m]
    i'm sure the user does not want this, in fact! :)
    breadhunter
    @EmiM
    This is the best I could provide at the moment: https://github.com/ZbayApp/waggle/runs/3705573421?check_suite_focus=true Historical run with old orbitdb.
    anders
    @tabcat00:matrix.org
    [m]
    this should link all the changes made to replication since then orbitdb/orbit-db#933
    the change affecting the chunking of replicated events was probably part of orbitdb/orbit-db-store#117
    anders
    @tabcat00:matrix.org
    [m]
    ill finish orbitdb/orbit-db#937 today and see if haad wants to help me get replication chunking again. i dont want you guys to have to dig through that code and fix it yourselves.
    breadhunter
    @EmiM
    Thank you :)
    haad
    @haad:matrix.org
    [m]
    hey, thanks for the pings. regarding the replicated event: indeed, this is now sent only when the db has been (fully) replicated. in fact, the previous behaviour wasn't intended (replicated events "chunked"). however, if I understand the need described above, you want to be able to "know when some progress of replication has been made" and the replicate.progress event will give you that (see https://github.com/orbitdb/orbit-db-store/blob/main/src/Store.js#L100 for the callback signature and params). it's emitted for each replicated db log entry. with that, you can eg. show a nice progress bar, or hook to it to show the replicated data to the user. note though that when replicate.progress is fired, the DB index hasn't been updated yet with the replicated data.
    Holmes Wilson
    @holmesworcester:matrix.org
    [m]
    Thanks! @EmiM is this what we need?
    Holmes Wilson
    @holmesworcester:matrix.org
    [m]
    haad: we were also wondering about expected improvements in sync time. do you have any tests that show improvements in sync time in orbitdb 0.28.0? we'll prepare more detailed data on this but our tests show little or no change.
    So I'm wondering if perhaps there's some remaining bottleneck or bug that is resulting in expected performance gains not being achieved in practice.
    And thanks so much for your work on this!
    breadhunter
    @EmiM
    Thank you @haad:matrix.org for clearing this up. Yeah, we were relying on replicated instead of replicate.progress exactly for the reason (db index not being updated) but maybe we can switch to replicate.progress.
    Holmes Wilson
    @holmesworcester:matrix.org
    [m]
    Will there be a way to display an entry before sync is complete?
    anders
    @tabcat00:matrix.org
    [m]
    yes but its probably still best to just wait for the sync to complete in most cases. in real world scenario you would get smaller updates compared to 1000+ entries at once unless coming back from being offline for a while.
    i kind of agree with haad's expectated behavior now
    Holmes Wilson
    @holmesworcester:matrix.org
    [m]
    We created an issue for sync performance just to track it. Should we be seeing performance increases? Does 20-150 seconds for 1000 entries seem reasonable? Does it make sense that there hasn't been any improvement after the performance related changes in 0.28.0? orbitdb/orbit-db#938
    vinkabuki
    @vinkabuki
    Hello, I have an issue with closing orbitdb - it hangs (not always and only for one instance) or throws error and hangs (always for more than one instance). I've created a minimal repo, so you can see it yourself. It doesn't throw en error if you don't initialize keyvaluestore - sometghing with ipfs pubsub I guess, but maybe there is something wrong with mu setup, I was trying to close and even drop those kvstores, but it didn't work - https://github.com/vinkabuki/closeorbitdb
    Qi Li
    @aaronli:matrix.org
    [m]
    Hi guys. I am playing around with the OrbidDB field-manuel, and I am stuck in this section (https://github.com/orbitdb/field-manual/blob/main/01_Tutorial/04_P2P_Part_1.md#getting-a-list-of-connected-peers). The problem is when I print my node's swarm list, it doesn't show any /p2p-circuit/ipfs/xxxxx swarm. It only shows the ip4/ip_address/tcp/port_number. In this case, I am not able to use the defined function connectToPeer() to connect this peer.
    Does anyone come across this issue?
    Rafael Kazuo Sato SimiĆ£o
    @dntxos
    GM guys. Im trying to test orbitDB with current demos. But never connect peers... And not sync with another remote machina that i was testing... even on orbitdb control center i cant sync with