1. Komal Thareja

Komal Thareja

Forum Replies Created

Viewing 15 posts - 76 through 90 (of 557 total)
  • Author
    Posts
  • in reply to: file extension is changing when I download a file #9067
    Komal Thareja
    Moderator

      Hi Nirmala,

      I tried an example as below and it seems to be working:

      1st parameter is the local path where file should be saved and second parameter being the remote location of the file to be downloaded.

      node.download_file('hello.c','/home/rocky/hello.c')

      Could you please share which version of fablib are you using?

      Best,

      Komal

      in reply to: Performance Drop on ConnectX-6 After Release 1.9 #9042
      Komal Thareja
      Moderator

        Hi Rasman,

        I forgot to mention that the steps for installing iperf3 should be run as the root user. On your VM, I did the following:

        sudo su -
        
        curl -L https://github.com/esnet/iperf/releases/download/3.18/iperf-3.18.tar.gz > iperf-3.18.tar.gz
        tar -zxvf iperf-3.18.tar.gz
        cd iperf-3.18
        
        sudo apt update
        sudo apt install build-essential
        
        sudo ./configure; make; make install
        sudo ldconfig
        

        I also applied the following host tuning (node_tools/host_tune.sh) on both VMs:

        #!/bin/bash
        
        # Linux host tuning from https://fasterdata.es.net/host-tuning/linux/
        cat >> /etc/sysctl.conf <<EOL
        # allow testing with buffers up to 128MB
        net.core.rmem_max = 536870912
        net.core.wmem_max = 536870912
        # increase Linux autotuning TCP buffer limit to 64MB
        net.ipv4.tcp_rmem = 4096 87380 536870912
        net.ipv4.tcp_wmem = 4096 65536 536870912
        # recommended default congestion control is htcp or bbr
        net.ipv4.tcp_congestion_control = bbr
        # recommended for hosts with jumbo frames enabled
        net.ipv4.tcp_mtu_probing = 1
        # recommended to enable 'fair queueing'
        net.core.default_qdisc = fq
        #net.core.default_qdisc = fq_codel
        EOL
        
        sysctl --system
        
        # Turn on jumbo frames
        for dev in basename -a /sys/class/net/*; do
            ip link set dev $dev mtu 9000
        done
        

        With these changes, I’m now seeing bandwidth close to 10G (see snapshot below).

        Screenshot-2025-09-24-at-5.31.23-PM

        According to fablib.list_links(), links from GATECH are capped at 8G. I’d suggest trying a different site instead of GATECH.

        Screenshot-2025-09-24-at-5.41.06-PM

        Regarding the slice getting stuck at Submit: your keys may have expired. Please try running the notebook jupyter-examples-rel1.9.0/configure_and_validate/configure_an_validate.ipynb. This should automatically renew your keys if needed.

        If it still hangs at submit, please check /tmp/fablib/fablib.log for errors and share here.

        Best,

        Komal

        in reply to: Performance Drop on ConnectX-6 After Release 1.9 #9032
        Komal Thareja
        Moderator

          Hi Rasman,

          By default, the standard iperf3 version does not perform well with multiple streams. ESnet provides a patched version that resolves this issue and delivers significantly better performance. This fixed iperf3 is already packaged inside the container.

          If you would like to run it directly on the host, you can install it with the following steps:

          curl -L https://github.com/esnet/iperf/releases/download/3.18/iperf-3.18.tar.gz > iperf-3.18.tar.gz
          tar -zxvf iperf-3.18.tar.gz
          cd iperf-3.18
          sudo apt update
          sudo apt install build-essential
          sudo ./configure
          make
          sudo make install
          

          Additionally, please make sure that the script node_tools/host_tune.sh (included with the notebook) has been executed on the relevant nodes.

          If you continue to see lower bandwidth, kindly share your slice ID so I can take a closer look.

          Thanks,
          Komal

          in reply to: can’t see nvidia card though VM shows component assigned #9025
          Komal Thareja
          Moderator

            Thank you for reporting this Maureen! We have identified the issue and working to identify a solution.

            We will keep you posted about the resolution. Apologies for the inconvenience.

            Best,

            Komal

            in reply to: Availability of DPU-powered SmartNICs #8989
            Komal Thareja
            Moderator

              Hi Tanay,

              We’re in the process of deploying them and are targeting DPU availability at KNIT11 around October 13–14.

              Best.

              Komal

              in reply to: Node Naming error #8931
              Komal Thareja
              Moderator

                Hi Nishanth,

                Could you please try again? It should work now, freebsd fix introduced this check, i have disabled it now.

                Best,

                Komal

                Komal Thareja
                Moderator

                  Hi,

                  Fix for this issue has been deployed on production. Please try creating a slice and let us know if you run into any issues.

                  Best,

                  Komal

                  in reply to: Establishing connection between different Slices. #8911
                  Komal Thareja
                  Moderator

                    Hi Tejas,

                    You can use FABRIC’s Layer 3 FabNetv4 or FabNetv6 Network Service to establish connectivity between slices.

                    Any VM connected to FabNetv* in one slice can communicate with a VM connected to FabNetv* in another slice, provided the routes are configured correctly. You just need to add the following routes:

                    ip route add 10.128.0.0/10 via <fabnetv4_gateway>
                    ip -6 route add 2602:FCFB:00::/40 via <fabnetv6_gateway>
                    

                    You may also find this example artifact helpful, as it demonstrates inter-slice connectivity using FabNetv4.

                    Best,

                    Komal

                    Komal Thareja
                    Moderator

                      Thank you for sharing this, Nishant and YoursSunny.

                      I was able to reproduce the issue. On IPv4 sites, user SSH keys are not being injected, and on IPv6, SSH connections are failing completely. We’ll work on addressing this and will let you know once the fix has been deployed. Apologies for the inconvenience in the meantime.

                      Best,

                      Komal

                      in reply to: Internal Server Error when running JupyterHub cell #8883
                      Komal Thareja
                      Moderator

                        Hi Dagim,

                        Thank you for sharing this observation. Could you update the instantiation of the fablib object in the first cell to the following and then try running the notebook again?

                        fablib = fablib_manager(project_id=project_id, validate_config=False)
                        

                        Thanks,
                        Komal

                         

                        Komal Thareja
                        Moderator

                          Hi Zhihe,

                          This was a bug, a fix has been deployed on the default container. Could you please try running this notebook again?

                          Thanks,

                          Komal

                          in reply to: earlier versions of Jupyter examples #8877
                          Komal Thareja
                          Moderator

                            Yes please Nirmala – Just keep the following in that file:

                            This should allow you to delete older examples.

                            {
                            "examples": [
                            {
                            "url": "default",
                            "location": "/home/fabric/work"
                            }
                            ]
                            }

                            Best,

                            Komal

                            Komal Thareja
                            Moderator

                              Hi Philip,

                              We discussed this internally, and the minimum supported bandwidth value is 1 Gbps. At this time, we don’t have plans to provide more fine-grained options. If you need to simulate lower bandwidth, we recommend using tools such as tc.

                              Best,

                              Komal

                              in reply to: Energy monitoring of an allocation #8870
                              Komal Thareja
                              Moderator

                                Hi Jacob,

                                Thank you for sharing the details. I discussed the energy consumption measurement topic with the team earlier today.

                                As mentioned before, we do not currently support energy consumption measurements on the VMs. The network team also confirmed that such measurement capabilities are not available on the network devices.

                                For identifying the geolocation of hops, we recommend using the L2PTP (Layer 2 Point-to-Point) network, where the user explicitly defines the network path by specifying the SITEs for each hop. You can then use fablib.list_sites() to obtain the geo-coordinates.

                                Please refer to this artifact for guidance on setting up an L2PTP slice:
                                https://artifacts.fabric-testbed.net/artifacts/7e439627-96be-45e0-ab67-50bb607f06e4

                                 

                                Also, regarding the renewal request, I see that it was submitted yesterday. Michael will follow up on that through the ticket.

                                Best,

                                Komal

                                in reply to: earlier versions of Jupyter examples #8868
                                Komal Thareja
                                Moderator

                                  Hi Nirmala,

                                  You don’t need to keep the example notebooks, so please go ahead and remove them.

                                  Could you also take a look at the contents of /home/fabric/fabric_config/fabric_config.json? I suspect there may be multiple entries in that file. If so, please delete it as well—this should prevent the older examples from being retained when you log in again.

                                  Best,

                                  Komal

                                Viewing 15 posts - 76 through 90 (of 557 total)