by

Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
    Mark Robert Henderson
    @aphelionz
    yeah so far so good with codesandbox! Thank you for the suggestion :)
    anders
    @tabcat
    no problem!
    Mark Robert Henderson
    @aphelionz
    made an orbitdb boilerplate so we can share our code ideas
    Mark Robert Henderson
    @aphelionz
    Rekindling and pinning this discussion re: monorepo: orbitdb/orbit-db#320
    Please vote and share your thoughts
    vaultec
    @vaultec81
    Cool I love the code sandbox template idea. Thanks @aphelionz
    Mark Robert Henderson
    @aphelionz
    Np
    phillmac
    @phillmac
    @aphelionz Someone had a PR for an internal operations queue for this type of error I'm sure of it
    anders
    @tabcat
    what type of error
    phillmac
    @phillmac
    phillmac
    @phillmac
    @vaultec81 Was That PR orbitdb/orbit-db-store#85 mentioned as a candidate in the draft PR?? I feel like it was but I can't remember
    anders
    @tabcat
    im unfamiliar with how the http-api works but the issue, which like you suggested may be caused by concurrent _addOperation calls, might be remedied by adding the queue or changing how _addOperation is called to a series
    phillmac
    @phillmac
    Its an async http api server thats just calling await db.put(...)
    anders
    @tabcat
    the guy with the issue might just be calling db.put multiple times without waiting for the one before to complete
    phillmac
    @phillmac
    yeah he's load testing, im guessing simulating multiple concurent clients
    imho its a valid use case, just not one that I designed for or ever tested XD
    vaultec
    @vaultec81
    @phillmac Yeah it was mentioned in the draft release orbitdb/orbit-db-store#85
    anders
    @tabcat
    is the benchmark-replication broken for anyone?
    oh i think im supposed to add an address as an argument, nvm
    Haad
    @haadcode

    Looking at orbit-db-store - I don't see any use of refs. Any downsides to using refs along with nexts here: https://github.com/orbitdb/orbit-db-store/blob/master/src/Replicator.js#L181

    @mistakia I believe that has been an oversight and we didn't add the usage of refs when we added those to ipfs-log. I've found some problems with that recently and it looks to me we should do a proper refactoring of Replicator (as ipfs-log and Replicator do same/similar things), as well as fix that lack of refs as it directly effects the replication throughput/performance (quite heavily based on my tests).

    @haadcode what's this for? is it redundant? https://github.com/orbitdb/orbit-db-docstore/blob/master/src/DocumentStore.js#L6 I get 'Readable' is assigned a value but never used

    @phillmac I have no idea where that comes from or why it's there. sounds like we should remove it?

    Haad
    @haadcode

    @haadcode I am just trying to understand OrbitDB and its limits right now. Is it true that the time to load OrbitDB is therefore proportional to the number of changes (inserts/updates/deletes) made to it? Or is there some kind of checkpoint where the whole db is flushed to avoid having to read the entire log to determine the current state?

    @chafey there's currently no functionality for "just give me the state of the DB" without reading the log. such state snapshotting would need to be implemented on the store level, prolly specific to each store. in addition, verifying that the state is correct, in multi-user/collaborative/trustless setting, would still prolly require to load the log (at least partially). one possible workaround would be to somehow save the DB state, say as a JSON object to IPFS, add the hash of the state to another log/feed db and upon loading the original db, load the state snapshot first from the db it was saved to. that said, I'm unsure if all that would be supported out of the box.

    gasparyanyur
    @gasparyanyur
    Hi all. I am getting an error look like this
    unable to save log data on store: cannot access config, repo not open
    I am using https://github.com/berty/go-orbit-db
    Can anybody help to solve this error ?
    Mark Robert Henderson
    @aphelionz
    Community call soon, on the hour orbitdb/welcome#91
    Mark Robert Henderson
    @aphelionz
    @gasparyanyur I know some folks from Berty were on here but it might be better to open an issue there
    ping @glouvigny
    vaultec
    @vaultec81
    Has anyone here heard of gundb before? https://gun.eco/docs/API
    Mark Robert Henderson
    @aphelionz
    Yeah we have
    vasa
    @vasa-develop

    Hey!

    What are the possible workarounds for the issue explained in this image

    issue.png
    context: 3Box uses orbitdb
    vaultec
    @vaultec81
    Is the slowness referring to read or write operations?
    vasa
    @vasa-develop
    yup...I think so
    If that is not the case, I can ask the person who wrote the above text.
    Mark Robert Henderson
    @aphelionz
    Is it write or replicate?
    Hmm I see
    What kind of DB is it? Which store type?
    vasa
    @vasa-develop
    let me see
    Kia
    @mistakia
    I've been struggling with peer discovery on the latest version of IPFS when it comes to peers behind NATs. It doesn't seem like websocket-star has been upgraded to the latest version of libp2p and I can't get libp2p-stardust to work. Would appreciate any pointers in the right direction.
    Mark Robert Henderson
    @aphelionz
    Oh jeez, I havent played with this but have been keeping an eye on it. If you can find @hugomrdias, @vmx, or @alanshaw on discord / matrix, they can at least point you in the right direction.
    You're free to tell them I sent you
    at least this one time, you can :D
    Kia
    @mistakia
    cheers - I think I actually have it sorted. Looks like they are about to finalize and push js-libp2p-stardust, just caught them at a bad time :D
    Mark Robert Henderson
    @aphelionz
    ah ok
    well, good!
    Mark Robert Henderson
    @aphelionz
    @tabcat I think the queue operations might have broken these tests, can you have a look?
    image.png