ETM | At first , ETM need load word embeddings | Topic Modeling library

I have a playbook that is launched from AWX with machine credentials, ie. with ssh_user_A. In this playbook, i need to perform a couple of tasks with another SSH user, ie. ssh_user_B.

In terms of credentials, I have:

• machine credentials for ssh_user_A that are OK, I can do whatever I need
• a public/private keys couple for ssh_user_B:
  • the public key is present on the remote host in the ~ssh_user_B/.ssh/authorized_keys file
  • if I test this public/private key with a temp machine credentials (not possible in final target) and a dummy playbook or an ad-hoc module call, it works I can do whatever I need

I change the SSH user in the playbook with these instructions:

I have such enum/range:

period: 'UTD' | 'EOM' | 'ETM'

I want to iterate through all options in select:

I have ec2 instance with ubuntu v20.04 and it has python v3.8.10 and pysftp 0.2.9.

I have generate .pem file from .ppk file using below command

puttygen sftp_server.ppk -O private-openssh -o sftp_server.pem

I am able to connect successfully to sftp server using command line-

I have a Python 3 application running on CentOS Linux 7.7 executing SSH commands against remote hosts. It works properly but today I encountered an odd error executing a command against a "new" remote server (server based on RHEL 6.10):

encountered RSA key, expected OPENSSH key

Executing the same command from the system shell (using the same private key of course) works perfectly fine.

On the remote server I discovered in /var/log/secure that when SSH connection and commands are issued from the source server with Python (using Paramiko) sshd complains about unsupported public key algorithm:

userauth_pubkey: unsupported public key algorithm: rsa-sha2-512

Note that target servers with higher RHEL/CentOS like 7.x don't encounter the issue.

It seems like Paramiko picks/offers the wrong algorithm when negotiating with the remote server when on the contrary SSH shell performs the negotiation properly in the context of this "old" target server. How to get the Python program to work as expected?

Python code

I am trying to use pip install git+ssh://git@bitbucket.org/my_org/my_package_repo.git to install a custom-made python package (shared by multiple applications) from BitBucket WITHOUT having to enter the SSH password.

There seems to be a lot of good information in one of the answers to this question for doing this in GitLab, etc. There's also some solid supporting information here.

I've already setup an SSH key between my local Mac and this BitBucket account. I'm able to push/pull code all the time to/from this account without having to re-enter the SSH password. Why is the pip install command requiring the password, when it's not required by git commands? Is there a way around this with BitBucket and the setup I've described?

When I run the GIT_SSH_COMMAND='ssh -vvv' pip install git+ssh://git@bitbucket.org/my_org/my_package_repo.git command recommended in the comments, I get the following (sanitized) response:

I am very new to AWS ec2, using linux, and using scp commands, just to get that out of the way. I have an AWS ec2 instance set up and want to upload a python file to that instance using an scp server command. The command I am using is scp -i C:\Users\Mason Choi\django_market_predict\Anchor_training_key_pair.pem C:\Users\Mason Choi\django_market_predict\mysite\train.py ec2-user@ec2-54-83-108-233.compute-1.amazonaws.com:/home/ec2-user

The error message I get when trying to use this command in my terminal (windows cmd) is

My scenario is that I have a Spring Integration Flow with a SftpInboundAdapter which shall get a file from a SftpServer "myHost". The file contains JSON which is to be converted into MyEvent entities to be forwarded to further processing. The process is implemented in a task scheduled by Spring Scheduler. Therefore the Integration Flow is not to start automatically with the Application having that autoStartup(false).

The Spring Integration Flow is:

• testSftpSessionFactory: to provide the session to the SFTP server
• testSftpInboundAdapter: to get the SFTP remote file
• sftpInputChannel: a Publish-Subscribe channel to have multiple message consumers
• sftpInputChannel-MessageHandler: to get the JSON content of the file transformed
• deleteLocalFileService-MessageHandler: to delete the remote file after successful processing
• controlChannel: to send Integration Flow control commands
• controlChannel-ExpressionControlBusFactoryBean: to start the testSftpInboundAdapter

The types TransferContext, TransferChannel and MyService are Java classes of mine with some fields fed from YAML properties which provide values for the sftpSessionFactory as host, port, user, password and the sftpInboundAdapter as remoteDirectory, remoteFilename, preserveTimestamp, localDirectory, etc. The Service is to process the MyEvent entities.

These are my SI beans:

I added following MACs to /etc/ssh/sshd_config of Ubuntu 18.04 compute instance on GCP. But after updating the file ssh is not restarting and journalctl -xe shows /etc/ssh/sshd_config line 130: Bad SSH2 mac spec.

I was using the exact private key sucessfully for 2 weeks to ssh into my EC2 instance. I tried today (without any changes from my side) and it is not working anymore...

Full verbose logs are as below: (where replaced all ip occurences with $IP) Facts:

1. ip is up and running, webserver is returning serving the page fine so instance is not stopped
2. i made no manual changes in last 2 weeks to cause this issue
3. Before 2 months I changed the ssh key by adding in known keys in ec2 instance the current key. I added it to user ubuntu at .ssh/authorized_keys. It worked correctly for 3 weeks and since then instance is not restarted/stopped