Trying to install Enterprise Version, failing at sshpass tool not found

TigerGraph DB
#1

Hi,

I am trying to install TigerGraph with a cluster option and its failing for me at the below-mentioned stage. The machine has the required software including sshpass.

[ERROR ]: Cannot find the tool(s): sshpass

Am I missing something or do I need to make any settings change for this?

TIA,

Anoop

#2

Hi Anoop,

Can you please try to run install_tools.sh after you get missing tool error and rerun install.sh again?

Thanks,
Chengbiao

#3

Thanks for getting back Checgbiao.

I have tried the steps you have mentioned and here is the response from install_tools.sh

[ERROR ]: No miss tools found, file utils/miss_tools NOT exist, exit

but I am still getting the same error. Can you please guide?

Thanks,
Anoop

#4

Hi Anoop,

In this case, please make sure sshpass is in the PATH environment variable if it’s already installed.

You can run “which sshpass” to verify, which should return the full path to the sshpass tool.

Thanks,
Chengbiao

#5

It is returning the full path in this case.

[root@stphub01 tigergraph-2.5.0-offline]# which sshpass

/usr/local/bin/sshpass

#6

OK. So you’re using root user. Is it the user you configured as “sudo” user for installer?

Thanks,
Chengbiao

#7

Yes, I configured “sudo” user for the installer as well.

#8

That is weird… Is it OK to provide the whole screen output start from executing the install.sh? Please do “echo $PATH” before start the installer and include it as well.

#9

Hi Chengbiao,

Please find attached the PATH and full installation logs. I am removing the IP addresses and the LIC information.

Thanks,

Anoop

#10

Can you try to launch the installer without sudo, i.e. run "./install.sh -c -l xxxxxxxx” directly?

#11

OK. So this worked. Thanks!

Now, what will be the next steps. It left me at a message: login into the machine.

I am logged into this machine, I dont know what the next steps are going to be because as per the documentation I should be able to execute the below mentioned steps but when I tried “gadmin” it actually gave me “bash: gadmin: command not found”. Can you please help me with that as well?

  1. The installer concludes by prompting the user to login to node m1 of the cluster and use ‘su’ to switch to the tigergraph user account. To confirm correct operation:
  2. Try the command gadmin status from any machine in the cluster.If the system installed correctly, the command should report that zk , kafka , dict, nginx, gsql, and Visualization are up and ready. Since there is no graph data loaded yet, gse , gpe , and restpp are not initialized.
  3. Try the command gsql --version The gsql command must be run on node m1 of the cluster because the gsql server is installed on m1 only. ``
  4. Basic installation is now finished! Please see Post-Installation Notes below.
#12

Did installation finished successfully? If yes, try to login with user “tigergraph” and try the gadmin commands, or do “su - tigergraph” to switch to the user from root.

Thanks,
Chengbiao

#13

Thanks Chengbiao,

your things worked but there is still one thing. I think its not installed properly as when I tried running gsql on the second node, its giving me that dev is not installed whereas on the first node I was able to execute the commands without any issues.

How will I be able to check if the installation was successful and replication is happening?

Thanks,

Anoop

#14

Hi Anoop,

Gsql is on the 1st node only hence the behavior is expected. The endpoint of the queries can be accessed from both machines though.

You can check system status with “gadmin status” and make sure every service is up.

Thanks,
Chengbiao