FAILED - RETRYING: wait for hosts (12 retries left) about sap-deployment-automation HOT 9 CLOSED

Hi @vasu2809, the error "ssh: connect to host 10.128.10.7 port 22: Operation timed out" is typically caused by a firewall blocking access, so I am wondering if you see the same timeout when ssh-ing manually from the same host where you are running Ansible. If it were an issue with the user not being present, you would more likely see a permission denied error rather than a timeout.

from sap-deployment-automation.

Thanks @rjosephwright - I am running Ansible from my Mac desktop (Visual Studio) and the SSH port is right now open to 0.0.0.0/0, I am also thinking if IAP is an issue here - but if that's the case, I believe IAP problem should come when I try to SSH from GCP console as well, I will keep digging though

from sap-deployment-automation.

IAP isn't used here by Ansible, so the ssh connection from the GCP console happens in a different way. I think the problem is that Ansible connects to the machines on their internal IP, and you are running Ansible from your desktop which cannot connect to them (unless you have e.g. Cloud Interconnect, but it sounds like that is not the case). I recommend running Ansible from a machine in the VPC to have direct access.

from sap-deployment-automation.

I tested the same script through AWX tower - installed it using the image available in the same network in which I am trying to install SAP application.

I get the below error while trying to deploy S4HANA-HA stack - now that AWX is able to SSH into the instances, I think it is not able to install the packages I believe and fails after 50 retries -

Wondering if we are missing any IAP configurations here -

{
"msg": "No provider of '+unrar' found.",
"rc": 104,
"stdout": "\n\n<message type="error">Package '+krb5-32bit' not found.\n<message type="error">Package '+libstdc++6-32bit' not found.\n<message type="error">Package '+pam-32bit' not found.\n<message type="error">Package '+python3-pip' not found.\n<message type="error">No provider of '+numactl' found.\n<message type="error">No provider of '+python-pip' found.\n<message type="error">No provider of '+unrar' found.\n\n",
"stderr": "",
"cmd": [
"/usr/bin/zypper",
"--quiet",
"--non-interactive",
"--xmlout",
"install",
"--type",
"package",
"--auto-agree-with-licenses",
"--no-recommends",
"--",
"+krb5-32bit",
"+unrar",
"+numactl",
"+python-pip",
"+python3-pip",
"+libstdc++6-32bit",
"+pam-32bit"
],
"invocation": {
"module_args": {
"name": [
"libopenssl1_1",
"tuned",
"krb5-32bit",
"unrar",
"numactl",
"sap-suse-cluster-connector",
"socat",
"libgcc_s1",
"libstdc++6",
"libatomic1",
"python-pip",
"python3-pip",
"libstdc++6-32bit",
"pam-32bit"
],
"state": "present",
"type": "package",
"disable_gpg_check": false,
"disable_recommends": true,
"force": false,
"update_cache": false,
"oldpackage": false,
"extra_args_precommand": null,
"extra_args": null
}
},
"stdout_lines": [
"",
"",
"<message type="error">Package '+krb5-32bit' not found.",
"<message type="error">Package '+libstdc++6-32bit' not found.",
"<message type="error">Package '+pam-32bit' not found.",
"<message type="error">Package '+python3-pip' not found.",
"<message type="error">No provider of '+numactl' found.",
"<message type="error">No provider of '+python-pip' found.",
"<message type="error">No provider of '+unrar' found.",
""
],
"stderr_lines": [],
"_ansible_no_log": false,
"attempts": 50,
"changed": false
}

from sap-deployment-automation.

The only time I have seen this error was when the Suse cloud registration failed. Some versions of Suse are more susceptible to this than others. You can verify it in the machine by running SUSEConnect -s and it should show that the system is registered for all of the available products. If it is not, you can run registercloudguest --force-new, or delete the failed machines and rerun.

from sap-deployment-automation.

Thank you - makes sense but all of the VM deployments for SAP HANA and Applications are deployed without internet access - will we still able to do the registration without network access for these VMs?

So, this would make NAT gateway deployment mandatory for the script to go through?

Resolution
Enable Internet access from the virtual machines to the Public Cloud Update Infrastructure servers. Once the virtual machine has network connectivity to the Public Cloud Update Infrastructure, you can have the on-demand virtual machine register to the SUSE Public Cloud Update Infrastructure by executing the following command as root:

registercloudguest --force-new

https://www.suse.com/support/kb/doc/?id=000019476

from sap-deployment-automation.

Doing a deployment without internet access requires more effort and will probably require building a custom image. If you just want to do a POC, I would highly recommend adding the NAT gateway.

from sap-deployment-automation.

Thank you for the prompt responses @rjosephwright - so the question is if we roll this out to customers - can we have NAT gateway listed out as one of the prerequisites to be made available in the network or maybe bake it in the deploy-vars.yml file unless a SUSE for SAP image with the packages is made available for use.

Also, I enabled NAT gateway, ran registercloudguest --force-new on all instances of HANA and SAP and also ran python3 installation on all of them manually -and the script through AWX tower is going through now.

If I may point out a couple of things I noticed while executing script from my desktop (Visual Studio) and through AWX tower -

1. AWX prompted for sap_hana_instance_name character limit to 7 - which wasn't the case with my Visual Studio, I was able to provide more than 10 characters and the instance was created fine
2. Can we also have a dropdown input for sap_nw_instance_type for choosing the instance type just like how we have for HANA?
3. The script did not create the Service Account automatically when it doesn't exist- it has to be created manually else script threw an error.

from sap-deployment-automation.

@rjosephwright Thank you, I have tested most of the scripts and everything looks good, the only intermittent issue I see is with python installation package, and even that happens only once in a while, if I retry the same script without performing any additional manual steps like registercloudguest --force-new, it goes through fine.

from sap-deployment-automation.