Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
    Luca Cominardi
    @Mallets
    Hi @geoff_or:matrix.org , zenoh uses shared memory if and only if the application has been programmed to make use of SHM. Otherwise it does not automatically translate into SHM.
    Geoff
    @geoff_or:matrix.org
    [m]
    Hmm OK. I wonder where all those DATA messages have gone then...
    4 replies
    I'll keep hunting
    Luca Cominardi
    @Mallets
    The nodes on the same host make use of the same zenoh session? I.e., if they belong to the same process and share the same session object, in that case zenoh will route data internally without sending them on the network.
    geoff_or
    @geoff_or:matrix.org
    [m]
    No, separate processes
    Luca Cominardi
    @Mallets
    You don’t see those messages in wireshark or you don’t see them neither at the API?
    another thing: are you running in peer-to-peer?
    if that’s the case, the zenoh peer running on the same physical machine may open connections on ports different from the 7447 (since they are sharing the same IP/port space). In that case, traffic might go to a different port
    Geoff
    @geoff_or:matrix.org
    [m]
    When are LinkStateList messages sent? I'm seeing a huge number of them being sent during a session - 13562 during a 15-second session with 20 nodes.
    3 replies
    Geoff
    @geoff_or:matrix.org
    [m]

    I'm continuing to see weird message counts. This gist is the output from a run with 20 processes all on the same computer (but actual individual processes), running until all the nodes have started up and then running for 10 seconds.

    https://gist.github.com/gbiggs/ba538077768a4c79984aa6f8c7cc313a

    The first curiosity is the huge number of error messages that appear. These mainly appear during the start-up phase. I don't understand what they mean, but my best guess is that some nodes are sharing linkstate information with other nodes before those nodes are ready?

    The second is the number of LINKSTATEINFO messages. It's more than half of the total Zenoh messages found.

    It would be great if someone here who knows how to fix these issues could go through them in real-time with me some time today. I'm not sure what time the developers are available, but I'm guessing European time?

    Geoff
    @geoff_or:matrix.org
    [m]

    :point_up: Edit: I'm continuing to see weird message counts. This gist is the output from a run with 20 processes all on the same computer (but actual individual processes), running until all the nodes have started up and then running for 10 seconds.

    https://gist.github.com/gbiggs/ba538077768a4c79984aa6f8c7cc313a

    The first curiosity is the huge number of error messages that appear. These mainly appear during the start-up phase. I don't understand what they mean, but my best guess is that some nodes are sharing linkstate information with other nodes before those nodes are ready?

    The second is the number of LINKSTATEINFO messages. It's more than half of the total Zenoh messages found.

    It would be great if someone here who knows how to fix these issues could go through them in real-time with me some time today or tomorrow. I'm not sure what time the developers are available, but I'm guessing European time?

    kydos
    @kydos
    Hello Geoff, we can arrange that. I’ll follow-up via email.
    Geoff
    @geoff_or:matrix.org
    [m]
    I'm still using 0.5.0-beta.9. Changing to master and its new API is non-trivial so I haven't done it yet
    Francesco Pham
    @frankplus
    Hi, I'm working on the zenoh integration into the Yocto build system.
    I successfully was able to run version 0.5.0-beta.9 of zenoh and its C library.
    Due to how the Rust dependencies are managed in Yocto I wasn't able to run the latest dev version of Zenoh C library because it needs Zenoh to be released on crates.io
    So my question is: if and when will the 0.6.0-dev be released on crates.io?
    Geoff
    @geoff_or:matrix.org
    [m]
    In the new API of master, how can I create a custom value and publish it?
    in the previous API I created a ZBuf from a Vec<u8>
    Geoff
    @geoff_or:matrix.org
    [m]
    I'd like to put protobuf-serialised data, which is what I was previously using the ZBuf for
    Luca Cominardi
    @Mallets
    It would be enough to session.put(&key_expr, payload).await.unwrap(); where payload is a Vec<u8>. There is no need any more to expressely create a ZBuf, this is done automatically under the hood by zenoh.
    Geoff
    @geoff_or:matrix.org
    [m]
    cool thanks!
    Jimmy Johnson
    @catch-twenty-two
    I am new to rust and was thinking I would use the lapin library as a proof of concept to create a plugin to connect the Zenoh router to our Azure backend since it has a lot of support. It looks like it doesn't support amqp 1.0 which is what I need. Not really seeing a good alternative, anyone else have any suggestions?
    2 replies
    kydos
    @kydos
    Dear Zenohers I would like to share this news with you. This is an exiting development and we look forward in sharing more updates in the months to come.
    As a result of the ZettaScale spin-off (link above we will be expanding our R&D teams, thus we have several openings across our offices. If any of you interested in joining, please reach out. I’ll be posting job descriptions on linked-in in the days to come.
    Geoff
    @geoff_or:matrix.org
    [m]
    I've converted my application to use Zenoh main branch, but don't see any difference in behaviour regarding linkstatelist message counts and link state errors
    Andreas Lööw
    @icucode
    How can I make sure that a peer reconnects after a remote peer restarts? It doesn't seem to do that by default!?
    3 replies
    Carlos Guimarães
    @cguimaraes

    Hello zenohers! A new blog post unveiling our DragonBot (as a fully Zenoh-powered TurtleBot look-a-like) has just been published.

    https://zenoh.io/blog/2022-02-08-dragonbot/

    Enjoy the reading!

    2 replies
    kydos
    @kydos
    Hello Zenohers, we are starting to feel the limitations of Gitter in terms of rooms management, communication threads and more importantly the ability to organise live community events, screen sharing sessions, etc... As such we are considering to move to either Discord or Slack, with a slight inclination for the former. We are evaluating the two and will get back to you ASAP on our recommendation. Meawhile, please feel free to comment or suggest potential alternatives — FYI we’ve already looked into Elements.
    6 replies
    Geoff
    @geoff_or:matrix.org
    [m]
    I'd prefer Slack over Discord. It's got better instance management.
    expploitt
    @expploitt
    I'm agree with @geoff_or:matrix.org
    kydos
    @kydos
    Thanks for your feedback.
    Andreas Lööw
    @icucode
    Is it possible to change the zenoh-bridge to use unicast instead of multicast for local discovery? We have a problem with multicast discovery on localhost when running in a container environment.
    6 replies
    Andreas Lööw
    @icucode
    I noticed that when using the zenoh-bridge (DDS) already published messages (transient-local) from peers are lost during startup. I can see the remote message in the started nodes rest-api but the data is not relayed to the local DDS instance. Is this a known issue?
    gunwale
    @andreaz98:matrix.org
    [m]

    hello, while building zenoh on linux Ubuntu 20.04 64 bit I get the following error on the building of zenohd: error: could not compile zenohd

    Caused by:
    process didn't exit successfully: rustc --crate-name zenohd --edition=2018 zenohd/src/main.rs --error-format=json --json=diagnostic-rendered-ansi,future-incompat --crate-type bin --emit=dep-info,link -C opt-level=3 -C panic=abort -C lto=fat -C codegen-units=1 -C metadata=c9dda7dc810ab693 -C extra-filename=-c9dda7dc810ab693 --out-dir /home/ubuntu/Desktop/smart_home_project/zenoh/target/release/deps -L dependency=/home/ubuntu/Desktop/smart_home_project/zenoh/target/release/deps --extern async_std=/home/ubuntu/Desktop/smart_home_project/zenoh/target/release/deps/libasync_std-937200eeccdf2242.rlib --extern clap=/home/ubuntu/Desktop/smart_home_project/zenoh/target/release/deps/libclap-73a5e9f3a56291a3.rlib --extern env_logger=/home/ubuntu/Desktop/smart_home_project/zenoh/target/release/deps/libenv_logger-a3675ac161571566.rlib --extern futures=/home/ubuntu/Desktop/smart_home_project/zenoh/target/release/deps/libfutures-cf64b90bd61c6914.rlib --extern git_version=/home/ubuntu/Desktop/smart_home_project/zenoh/target/release/deps/libgit_version-2e04fa2a4f21e59f.rlib --extern json5=/home/ubuntu/Desktop/smart_home_project/zenoh/target/release/deps/libjson5-bddce3a5523bcd13.rlib --extern lazy_static=/home/ubuntu/Desktop/smart_home_project/zenoh/target/release/deps/liblazy_static-a47345916819c0a5.rlib --extern log=/home/ubuntu/Desktop/smart_home_project/zenoh/target/release/deps/liblog-6423d0c548cde02c.rlib --extern validated_struct=/home/ubuntu/Desktop/smart_home_project/zenoh/target/release/deps/libvalidated_struct-d1943142a9c15cbf.rlib --extern zenoh=/home/ubuntu/Desktop/smart_home_project/zenoh/target/release/deps/libzenoh-588010e32237ee8b.rlib -L native=/home/ubuntu/Desktop/smart_home_project/zenoh/target/release/build/ring-2993e85583e37710/out (signal: 9, SIGKILL: kill)
    warning: build failed, waiting for other jobs to finish...
    error: build failed

    2 replies
    ryan
    @ryan:matrix.hagen.tech
    [m]
    I'll be starting a project to build the router into a k8s application soon. Wanted to tap this channel for any tips before i get started.
    2 replies
    Charles Cross
    @spiderkeys_gitlab

    Hey Zenoh team! Working with pico a bit more, we've reached a point where we are interested in making sure that messages can be validated with a checksum. Of course, this can be done presently in two ways - potentially in the IP layer to validate individual packets (if used as a transport and supported as a feature), or by embedding a checksum in the custom encoded payload to validate the payload itself. However, it looks like there is no CRC in the current message protocol, at least within Pico, which would allow validation of all contents within a single zenoh message.

    Are there plans for this or perhaps existing ways to achieve it? Maybe via attachment decorators?

    11 replies
    ryan
    @ryan:matrix.hagen.tech
    [m]
    @Mallets: Just a router, yes it will need to be exposed outside the pod via a service. I will attempt this in azure Kubernetes service
    ryan
    @ryan:matrix.hagen.tech
    [m]
    :point_up: Edit: @Mallets: Just a router, yes it will need to be exposed outside the pod via a service. So it would be a client application, with the router in the cloud, and clients in the internet.
    2 replies
    :point_up: Edit: @Mallets: Just a router, yes it will need to be exposed outside the pod via a service. So it would be a client application, with the router in the cloud, and clients in the internet. We dont plan to have any peers.
    jpace121
    @jpace121:matrix.org
    [m]
    Hi! Are there any plans to add mutual SSL (aka client based) authentication to zenoh? I'm very interested in zenoh, but securing things with passwords with password databases I'd have to copy around is unappealing.
    2 replies
    Italkebby
    @Italkebby

    Hi, I would like to connect an ESP32 and a Rpi0W ver 2 and use ESP32 as zenoh client and RPi as a peer (router + client).
    So I try to install and run a zenoh router on Raspberry through docker, but after use this command:
    sudo docker run --init --net host eclipse/zenoh:master

    I received this error:

    WARNING: The requested image's platform (linux/amd64) does not match the detected host platform (linux/arm64/v8) and no specific platform was requested
    [FATAL tini (8)] exec /entrypoint.sh failed: Exec format error

    So is there a way to build a zenoh router for my CPU arch compatible with zenoh-pico version? Or is there a release in these months about a fresh version of zenoh-pico ?
    Thank you

    kydos
    @kydos
    @Italkebby one option is to build the zenoh master on your RasperryPI.the ESP32 running zenoh-pico (I suggest you also use the master), should work w/o any issue.
    2 replies
    Julien Enoch
    @JEnoch
    @Italkebby I managed to build an arm64 Docker image on Github. Please try docker pull jenoch/zenoh:docker_arm64 and let me know if it works on your Rpi0W.
    If it does, we’ll add it for the next release and possibly for nightly builds also (depending if we reach some Github limit, because this arm64 docker build takes 1h40!)
    11 replies
    kydos
    @kydos
    Dear Zenohers after some long discussion inside the team and feedback from the community we finally decided for the platform to host the community…
    Our shortlist was between Slack and Discord and in the end we decided to go with the latter as off the shelf it provides all we need, including the ability to do descktop sharing and video-conf as part of the open version.
    We have created a zenoh server which is available at https://discord.gg/cRFZDJfS3g and would like in due time to migrate the community there.
    I think the first compelling event is the zenoh workshop with the zenoh developer team we plan to stream in a couple of week — more to come on this.
    Andreas Lööw
    @icucode
    Testing with forward_discovery: true in the zenoh-dds-bridge seems to disable reconnect? If a node disconnects it seems like the bridge goes into an incorrect state (forward_discovery: false is working).
    13 replies
    robruh
    @robruh
    This message was deleted
    1 reply
    James Laney
    @jwrl_gitlab
    Is there a particular ordering that needs to happen when passing the various flags for the zenoh-plugin-dds?
    Iv notice that placing the flags in different orders may end up in not having an effect / enabling that particular setting.
    3 replies
    James Laney
    @jwrl_gitlab

    any insight on this error:
    [2022-03-28T21:50:26Z ERROR zplugin_dds::dds_mgt] ** END OF PERIODIC READER LOOP

    with more context:
    [2022-03-28T21:50:15Z INFO zplugin_dds] Remove unused route: DDS 'rq/gx5/ros2_mscl_node/get_available_transitionsRequest' => zenoh '/rq/gx5/ros2_mscl_node/get_available_transitionsRequest'
    [2022-03-28T21:50:16Z ERROR zplugin_dds::dds_mgt] END OF PERIODIC READER LOOP
    [2022-03-28T21:50:25Z INFO zplugin_dds] New route: DDS 'rq/gx5/ros2_mscl_node/get_available_transitionsRequest' => zenoh '/rq/gx5/ros2_mscl_node/get_available_transitionsRequest' with type 'lifecyclemsgs::srv::dds::GetAvailableTransitionsRequest'
    [2022-03-28T21:50:25Z INFO zplugin_dds] New route: zenoh '/rr/gx5/ros2_mscl_node/get_available_transitionsReply' => DDS 'rr/gx5/ros2_mscl_node/get_available_transitionsReply' with type 'lifecyclemsgs::srv::dds::GetAvailableTransitionsResponse'
    [2022-03-28T21:50:25Z INFO zplugin_dds] Remove unused route: zenoh '/rr/gx5/ros2_mscl_node/get_available_transitionsReply' => DDS 'rr/gx5/ros2_mscl_node/get_available_transitionsReply'
    [2022-03-28T21:50:25Z INFO zplugin_dds] Remove unused route: DDS 'rq/gx5/ros2_mscl_node/get_available_transitionsRequest' => zenoh '/rq/gx5/ros2_mscl_node/get_available_transitionsRequest'
    [2022-03-28T21:50:26Z ERROR zplugin_dds::dds_mgt]
    END OF PERIODIC READER LOOP

    When i start and stop a subscriber on the remote side this seems to pop up

    1 reply
    kydos
    @kydos
    Hello Zenohers, I am glad to see that many of you hav moved to the Zenoh Discord Server. For those that missed my last post, please consider joining our discord server at https://discordservers.com/server/914168414178779197