2 Replies Latest reply on Feb 12, 2019 5:54 AM by Yaron Lirase

    Slowness in copying admin bundle to worker ?

    Yaron Lirase

      Hello All,

       

      I am trying to launch a temporary distributed Tableau infrastructure on AWS.

      Unfortunately, always when I configure the worker, the process of copying the admin bundle (worker.zip) to the worker takes extremely long (over 10 hours).

       

      I have performed network tests between the Primary & Worker servers.

      By using iperf I see the network speed is very good (1.02 Gbits/sec).

      This test was also conducted on the ports used by Tableau for copying the files (ports: 51237, 5146 and just in case port 80).

       

      I can see that during this process, the network utilization on both Primary & Worker is at a bare (and so are other resources such as CPU, Memory etc.)

      Therefore it does not seem to me like a Network issue, but perhaps a problem with Tableau.

       

      Some details of the architecture:

      * Both EC2 instances are setup in the same availability zone, same SubnetId

      * All ports are open between the servers.

      * The hostnames are resolved to the VPC private IP addresses (to ensure the communication is done using the private IP only)

      * OS - Windows Server 2008 R2 (64 bit)

      * Worker has the following Tableau options selected: 2 Backgrounders, 2 Cache Servers & Gateway enabled

       

      Each time the Tableau Initialization Progress seems stuck in the step: "Sending admin bundle to worker #1".

      However if I check the folder C:\ProgramData\Tableau on the worker, I can see it is receiving the file worker.zip, but extremely slowly.

       

      Has anyone experiences a similar issue? I can provide more details if needed.

       

       

      Thank you in advance to any help,

       

      Yaron