cannot ssh vanilla aws instance - amazon-web-services

I've spend my whole evening on this. guess it is just I missed a stupid step. Here is the procedure I followed:
create an aws vpc 10.0.0.0/24;
create an aws internet gateway and associate it with the VPC;
create a subnet in the VPC 10.0.0.0/26;
Add inbound rule to VPC ACL to allow SSH, HTTP, HTTPS from all IPV4 sources;
Launch aws ec2 instance with Amazon Linux 2 AMI in region us-west-2, t2.micro,
instance details:
Number of instances: 1
network: VPC created above
subnet: subnet created above
auto-assign Public IP: use subnet setting(Disable)
Capacity reservation: Open
everything else as default
storage details
add on data volume, delete on termination check
security group: new security group with inbound rules ssh/http/https opened for all ipv4 sources
use existing key pair I created earlier;
create an elastic IP;
associate the elastic IP to the instance created above.
reboot the instance
Then I can see the instance is running well with elastic IP attached.
I tried to connect to the ip address with ssh ssh -vvv -i ./aws_private.pem ec2-user#ipaddress and got below failure
OpenSSH_8.1p1, LibreSSL 2.7.3
debug1: Reading configuration data /etc/ssh/ssh_config
debug1: /etc/ssh/ssh_config line 47: Applying options for *
debug2: resolve_canonicalize: hostname <elastic ipaddress> is address
debug2: ssh_connect_direct
debug1: Connecting to <elastic ipaddress> [<elastic ipaddress>] port 22.
debug1: connect to address <elastic ipaddress> port 22: Operation timed out
ssh: connect to host <elastic ipaddress> port 22: Operation timed out
I verified the ACL and Security group. Those are my only ACL and Security Group.
I'm using a Mac. And also tried with a windows OS as well as aws cloud shell in the same region. All of them have the same error.
telnet <elastic ip> 22 fails as well.
what did I do wrong?

You forgot one thing .
Security Groups
You need to allow SSH connection to your instance port 22. to do this add below rule to your Ec2 instance's security group.
Protocol - TCP, type- SSH , port -22, source - 0.0.0.0/0
Actually , you dont need to edit ACL , as those by default allow access to everything .
but you need to add rules to Security Groups as by default it Not allows any connection

In general, NACL rules should not be changed from default unless you have a very specific requirement (eg creating a DMZ).
NACLs are stateless, which means that they need to be specified in both directions. I recommend that you reset the NACLs back to their default setting of "Allow All" for both Inbound and Outbound, and then only use Security Groups for controlling access.
The flow of the connection will be:
Elastic IP address points to Internet Gateway
Internet Gateway does a reverse NAT to convert it to a private IP address
Route Table on Subnet is consulted to confirm that the subnet is "public" (Route Table entry pointing to Internet Gateway)
NACL checks traffic on entry to subnet (recommendation is to leave it as default Allow All)
Security Group checks traffic coming into the instance
Instance processes incoming request

turned out it was zscaler on my Mac somehow changes my ip address and messed up the communication between my Mac and aws.

Related

Not able to ssh/http into EC2 instance

I am at my wits end with this, please help.
I am creating EC2 instances in my default public VPC, yet i am not able to ssh or http to my instance or webserver running into the machine. I checked the following
The SG has inbound SSH, HTTP and HTTPS allowed from 0.0.0.0/0 and assigned to my instance
the default VPC, has route tables with 0.0.0.0/0 pointed to IGW
the NACLs are configured to Allow all traffic. i also manually updated to allow only HHTP, HTTPS and SSH
the Key is use has been given the right permission by running chmod 400 filename
Despite all this not able to connect to the EC2 instance, AMI being Amazon Linux 2 AMI
When I try to ssh, i get a connection timeout error after a while, initially, i thought it was my office network but I am also getting the same from my home network with no firewalls in place
To allow an SSH connection, you will need:
An Amazon EC2 instance running Linux launched in a public subnet (defined as having a Route Table that directs 0.0.0.0/0 to an Internet Gateway)
A Security Group permitting Inbound access on port 22 (Outbound configuration is irrelevant)
Network ACLs left at their default settings of Allow All in both directions
A Public IP address associated with the instance
From your descriptions, I would say that the problem is probably with the Outbound NACLs. Return traffic from an SSH session goes back to the source port on the initiating server, which is not port 22. In general, only change the NACLs if you have a specific reason, such as creating a DMZ. I recommend you reset the NACL rules to Allow All traffic in both directions.

Redshift ODBC Driver test fails: Is the server running on host and accepting TCP/IP connections on port 5439?

I have installed the 64x Amazon Redshift ODBC driver on my Windows 10 and it shows up under System DSN tab of ODBC Data Source Administrator (64-bit). I click on Configure and as I enter the connection information from my Amazon Redshift cluster properties, I face the following error:
Here are the current Network Security settings of the cluster. The only thing I have changed is that I have made it Publicly accessible (the value changed from No to Yes). But I haven't changed anything in VPC and VPC security group. As far as I see, all Inbound and Outbound traffic is allowed.
Anything I might have missed?
Should I add a firewall rule to my local machine for this port as well?
UPDATE:
Inbound rules for the VPC Group
Outbound rules for the VPC Group
If you check your security group (SG), you can notice that the source in the inbound rule is sg-9054xxxx. This means that all internet traffic is blocked.
You can only connect to your redshift from EC2 instance having same SG.
Alternatively, you can modify the source of the inbound rule in the SG to be your home/work ip address/range, or any IP address (not recommended).
Other option is to setup a bastion instance in a public subnet with ssh connectivity. Then you can setup ssh tunnel between between your windows and redshift using the bastion. Similar for VPN connection.
To query redshift you may also use its data api. This does not require establishing a regular connection to the cluster.

Aws Connection to EC2 timed out over SSH

I have tried to connect EC2 using SSH but ssh: connect to host XXXXXXXXX port 22: Connection timed out
Note: XXXXXXXX is user#IP
Also I have checked security groups. Inbound rules are allowed for ssh
SSH TCP 22 0.0.0.0/0 -
SSH TCP 22 ::/0 -
For first time, I was able to login using SSH. After that I installed LAMP stack on EC2 instance. I think I forgot to add ssh in ufw rules.
I can't able to connect using Browser Based SSH Connection in AWS and showing erros for Session Manager connection method.
How can I connect using SSH or other, so I can allow SSH in ufw rules.
This indicates that you cannot to the host.
From your question I can see you have validated security group access, however there are some other steps you should take to investigate this:
Is the IP address a public IP? If so ensure that the instances subnet has a route table with a Internet Gateway associated with it.
Is the IP address a private IP? Do you have a VPN or Direct Connect connection to the instance? If not you would either need to set this up or use a bastion host. Otherwise if you do ensure that the route tables reference back to your on premise network range.
If you're not using the default NACLs for your subnet check that both the port ranges from your security group as well as ephemeral port ranges.

Timeout SSH to public AWS VPC subnet (w/ public IP, routing table to igw, NACL allow all, security group allow SSH, etc.)

Very weird, have tried everything and no luck.
Internet Gateway attached to VPC.
A subnet is created with routing table 0.0.0.0/0 pointing to Internet Gateway.
Automatic assignment of public IP enabled.
NACL allows ALL traffic in and ALL traffic out on ALL ports.
Launched an EC2 instance in said subnet with a public IP and domain name.
Instance status is healthy.
Security group on instance allows SSH port 22 and PING from 0.0.0.0/0.
Flow Log on said subnet shows that inbound connection on port 22 is ACCEPT.
Ping public IP of EC2 instance receives responses (with loss of packets maybe due to region difference).
Still, cannot SSH into the instance! Have searched all over the net and manual and FAQ, no luck. Please help!
Additional info: The error was "Connection timeout", it responses that like 1-2 minutes after the ssh command.
SSH -v output (no useful info seems):
OpenSSH_7.6p1, LibreSSL 2.6.2
debug1: Reading configuration data /etc/ssh/ssh_config
debug1: /etc/ssh/ssh_config line 48: Applying options for *
debug1: Connecting to xxx.compute.amazonaws.com port 22.
ssh: connect to host xxx.compute.amazonaws.com port 22: Operation timed out
I tried associating an Elastic IP to the instance (keep all other settings untouched), SSH was successful. Then disassociate EIP and use old public IP, no luck. As per documentation it is not required to use Elastic IP for Internet access, auto-assigned public IP should also work. The same setting worked in the past with default subnets created automatically by AWS, which are public subnets and allowed SSH. Again, very weird!
The sad thing is I'm not sure how I could debug this situation and locate the real problem. Any help would be appreciated!

Accessing RDS through bastion host with port forwarding not working

I'm trying to establish a port forwarding to my RDS in a private subnet via a bastion host in a public subnet with the following command:
ssh -A -NL 3007:mydb3.co2qgzotzkku.eu-west-1.rds.amazonaws.com:3306 ubuntu#ec2-562243-250-177.eu-west-1.compute.amazonaws.com
but cant get a connection to the rds instance.
The security group for the Bastion Host allows only SSH on port 22 from my IP
and the security group for the RDS allows traffic from the bastion hosts security group and SSH from my iP
Besides the ACL for the subnets are open to all traffic for TCP.
anybody a tip what is missing to get the tunnel running?
merci A
I think you are missing the port 3306 and 3307. Allow that port in the both security group and it will work.
As you said you are accessing the bastion via key-pair, your new command must be:
ssh -N -L 3007:mydb3.co2qgzotzkku.eu-west-1.rds.amazonaws.com:3306 ubuntu#ec2-562243-250-177.eu-west-1.compute.amazonaws.com -i /path/to/key.pem
I would suggest removing A from the command as it Enables forwarding of the authentication agent connection. This can also be specified on a per-host basis in a configuration file.
Agent forwarding should be enabled with caution. Users with the ability to bypass file permissions on the remote host (for the agent's UNIX-domain socket) can access the local agent through the forwarded connection. An attacker cannot obtain key material from the agent, however they can perform operations on the keys that enable them to authenticate using the identities loaded into the agent.