I'm trying to access Tensorboard on AWS. Here is my setting :
Tensorboard : tensorboard --host 0.0.0.0 --logdir=train :
Starting TensorBoard b'39' on port 6006 (You can navigate to
http://172.31.18.170:6006)
AWS Security groups (in):
HTTPS TCP 443 0.0.0.0/0
Custom_TCP TCP 6006 0.0.0.0/0
However connecting to ec2-blabla.us-west-1.compute.amazonaws.com:6006 I can't see anything, I basically can't connect.
Do you have any idea?
You can use ssh tunneling technique.
In your terminal:
ssh -i /path/to/your/AWS/key/file -NL 6006:localhost:6006 user#host
where:
user and host: your aws ec2 user and instance specific.
-N: don't execute a remote command (just forward ports)
-L: [bind_address:]port:host:hostport
After that, browse to http://localhost:6006/
Run tensorboard in your ec2 terminal (you can custom logdir and port)
tensorboard --logdir=data/model --port=8080
Find your workstations public ip (a.b.c.d) address by visiting http://ip4.me/
Access the security group configuration assigned to your EC2 and add a custom TCP rule to your inbound traffic.
Outbound should be set to allow traffic from tensorboard port. (In this case 8080). Or you just allow all outgoing traffic from your EC2 instance
Protocol Port Range Destination Description
All traffic All All 0.0.0.0/0
Use your public DNS to access tensorboard from your workstation
http://ec2-xx-xxx-xx-xx.compute-1.amazonaws.com:8080/
Fast (but unsecure) solution:
Run:
tensorboard --logdir=/training --host=0.0.0.0 --port=8080
on your AWS instance.
Make sure that both your inbound and outbound rules on AWS console (control center) are as unrestricted as possible (allow all types, all ports etc.). However, keep in mind that this solution is not recommendable for environments requiring security (in our case, we didn't consider security for training an NN).
An attempt to explain why this works: when the policy is set as described, AWS still seems to prohibit inbound/outbound connections on the standard tensorboard port 6006. This does not seem to apply to the port 8080.
Long (but more secure) solution:
See: https://blog.altoros.com/getting-started-with-a-cpu-enabled-tensorflow-instance-on-aws.html
(provides explanations for setting ports correctly on AWS)
I managed to set it up like this:
Go to security groups in your ec2 console:
Choose the relevant security group in the table, click edit.
Add a rule like this:
Start tensorboard: tensorboard --logdir tf_summary/ --port 8080
Find out the URL of your instance and visit http://yourURL:8080
Simply run the tensorboard without the host parameter (which poses restrictions)
tensorboard --logdir XXX --port 6006
I suffered from the same problem for several days.
Fortunately I solved this issue by adding rule on "AWS Outbound rule" as if I had added "AWS Inbound rule".
Regardless of this setting, it works at home.
The same error is still happening only in the company.
Related
When i attempt to SSH from https://console.cloud.google.com/compute/instances?authuser=0&projec
I get the error:
Connection via Cloud Identity-Aware Proxy Failed
Code: 4010
Reason: destination read failed
Please ensure you can make a proper https connection to the IAP for
TCP hostname: https://tunnel.cloudproxy.app You may be able to connect
without using the Cloud Identity-Aware Proxy.
Though on a seperate machine i am able to login, but Firewall and other settings seem identical ?
What should i do to get in ?
Krgds.
C
This issue is due to the lack of a firewall rule to allow-ingress-from-iap with this IP range 35.235.240.0/20 that needs to be configured when using IAP (Identity-Aware Proxy).
To Allow SSH access to all VM instances in your network, do the following:
1- Open the Firewall Rules page (Navigation menu > VPC network > Firewall) and click Create firewall rule
2- Configure the following settings:
Name: allow-ingress-from-iap
Direction of traffic: Ingress
Target: All instances in the network
Source filter: IP ranges
Source IP ranges: 35.235.240.0/20
Protocols and ports: Select TCP and enter 22 to allow SSH
3- Click Create
My impression is that the root casuse, was the the linux box had run out of hard disk space, causing the issues.
So when i recreated the same machine, with more disk space, the issue was resolved.
I am getting the following disconnection issues in the GCP Jupiter notebook.
error code: 4010
and
error code: 1006
Can you please suggest some solution?
As part of the IAP configuration steps, you should create a firewall rule that allows ingress traffic to the SSH port from the IAP address range:
GCP Console => VPC network => Firewall rules => Create Firewall Rule
Name: allow-ingress-from-iap
Direction of traffic: Ingress
Target: All instances in the network
Source filter: IP ranges
Source IP ranges: 35.235.240.0/20
Protocols and ports: select TCP and enter 22 to allow SSH
Identity-Aware Proxy > Doc > Setting up IAP for Compute Engine
The error 1006 appears in the GCP Console UI after 1 hour of inactivity of the SSH session via IAP with VMs with Internal IP only, and this is a session timeout on the Google side.
As #mebius99 has mentioned, IAP (Identity-Aware Proxy) requests come from the IP address range 35.235.240.0/20.
Your network firewall must allow these requests to be able to SSH through IAP.
One way to do that (create a firewall-rule) is to run gcloud compute firewall-rules create command.
To do that, first open the cloud shell on the Google cloud console,
Then once the cloud shell opens up, run the following:
gcloud compute firewall-rules create ssh-ingress-from-iap --allow=tcp:22 --source-ranges 35.235.240.0/20 --network [network-name]
Replace [network-name] with your network name (the default VPC network is named: default)
If the above solution doesn't work (or have a similar firewall rule in place already), consider checking the network tags (on the firewall-rules and the VM). It maybe the case that your firewall-rule is allowing the requests to only certain instances that has some tags and the instance you're trying to SSH into doesn't.
For me the error 1006 was related to system Time. I had changed the system time manually to another time zone. SSH worked when system time was sy
Create a firewall for port 22 and add this IP 35.235.240.0/20
attach it to all VM so you will able to establish connection
Created an AWS AMI instance.
I can telnet from the instance itself
telnet [Pv4 Public IP] 9200
But not from my pc.
This is my security group
What am I doing wrong?
You can check your Network ACL configurations.
It looks like there is some other firewall in between your PC and server which is blocking you on 9200.
If you can access port 80 via telnet or you're able to SSH in it's likely you have a network ACL in place. If you can not access port 80 via telnet but you can via a browser it's like a local config - maybe AV or a firewall.
EC2 instances use security groups for their firewall
Another test to narrow down the the issue would to see if you could telnet from another instances in the same subenet in the same AZ. Being in the same subnet you should not be affected by a network ACL.
You can have a look at this telnet-to a cloud instance from outside
The solution to problem was "Open the services and make the telnet manual and right click on it and chose start"
As well make sure that the instance is residing in a public VPC
Based on what you've described, there isn't really much else to work with. Your ability to telnet the public IP from the instance implies the server is listening on the external interface and your security group is already set to have the port open to all incoming connections.
Aside from the trivial overlooking of not actually having the instance under the listed security group, the only possibility I can think of now is an active firewall on the instance. In the case of iptables or ufw (which is an interface to iptables), it's trivial to verify whether they are indeed getting in the way:
// List iptables access rules
sudo iptables -L -v
// List access rules via ufw
sudo ufw status
You said: "This is my SG", but...which way? Inbound or outbound?
It can simply be that your host can't reply to your PC.
Try to add a rule which adds outbound TCP ranging from ports 32768 to 65535 (ephemeral ports), so that the telnet server response packets can travel back to your PC.
Otherwise, like the others said, look at one level up, VPC-level (network ACL).
You might have your acceptor process running on 127.0.0.1:9000 which means only local clients can connect. This is not related to your Security Group which could be wide open.
Run lsof -i:9000 if on unix.
If you see something like this under NAME then host IP used to start your acceptor will needs to change from 127.0.0.1 to 0.0.0.0 (and secure via SG/FW).
COMMAND PID USER FD TYPE DEVICE SIZE/OFF NODE NAME
java 2777 ubuntu 148u IPv6 26856 0t0 TCP localhost:afs3-callback (LISTEN)
A Telnet service is not installed by default on an Amazon Linux AMI.
If you wish to use it, you will need to install it yourself, eg: Install and Setup Telnet on EC2 Amazon Linux or CentOS.
However, these days it is recommended to use ssh instead of telnet because it is more secure. See: Telnet on wikipedia
Just a thought, check firewall of your PC.
Need to ensure your SSH key you generated via IAM and attached to the EC2 at launch is added to the login:
ssh-add -K <yourkeyname>.pem
ssh ubuntu#<yourdns or ip>.com == or == ssh ec2-user#<yourdns or ip>
I am trying to access my ElasticSearch on a running EC2 instance from outside the Cloud. I currently have SSH/HTTP/HTTPS open to the public for inbound traffic as well as all open for outbound traffic. I set up a public IP for my EC2 instance as well.
By default ElasticSearch is on port 9200. I'm not sure if I configured my elasticsearch.yml file correctly but it basically has the default configuration I only changed the cluster.name to something else.
When I type in my public IP with port 9200 into my local browser or locally do a telnet {public-ip} 9200, there is no response. When I SSH into my EC2 instance. I can perform a curl localhost:9200 and I get the correct response from elasticsearch
How can I connect to my ElasticSearch running on my EC2 instance from outside the cloud?
I added a Custom Rule for my security group for inbound traffic that includes port 9200 and is open to 0.0.0.0/0 and I still cannot access this EC2 instance
Potential issues to check are wrong binding and instance operating system firewall.
Check where elasticsearch is binding, as if it is binding to 127.0.0.1 you won't be able to reach it from the outside.
Check binding by running in one shell on the elasticsearch ec2:
sudo netstat -lptun | grep 9200
If it shows 127.0.0.1:9200 then there is a misconfiguration if otherwise shows
*:9200 or :9200 then it is correct.
If it shows 127.0.0.1 then you should modify elasticsearch parameter network.bind_host as described in:https://www.elastic.co/guide/en/elasticsearch/reference/1.4/modules-network.html
Additionally http/HTTPS and ssh are usually allowed by default operating system firewall, whereas elasticsearch 9200 is not. This is usually the case for rhel and centos. You can temporarily disable iptables and check if it works.
To disable iptables run:
sudo iptables -F
If after disabling iptables the connection works you should configure iptables to allow connection on 9200.
I hope this helps.
G.
It is mess around Security Groups
You can add or remove rules for a security group (also referred to as
authorizing or revoking inbound or outbound access).
You shuld use the SG while launching your instance whith bounded 9200
Establish an SSH tunnel from your desktop to EC2.. then simply use your browser.. follow steps as given in https://www.jeremydaly.com/access-aws-vpc-based-elasticsearch-cluster-locally/
Network Newbie here.
Background
I am trying to use the AWS Command Line utility to connect to my instances on the AWS cloud.
Problem
When I try a simple command (from either local machine or EC2) such as...
aws ec2 describe-instances
I get...
Could not connect to the endpoint URL: "https://ec2.sa-east-1a.amazonaws.com/"
Which is fine and dandy because according to the AWS Command Line User Manual I should take note of:
The AWS CLI makes API calls to services over HTTPS. Outbound connections
on TCP port 443 must be enabled in order to perform calls.
But what that is suppose to mean in the first place? I find the above statement as clear as a glass of muddy water.
1) Should I go to the 'Security Group' of my instance and set inbound rules to 'HTTPS TCP 443 0.0.0.0/0'? (Doesn't work)
2) Or set outbound rules and set 'HTTPS TCP 443 0.0.0.0/0' ? (Doesn't work)
3) Does the above statement has anything to do with the 'Security Groups'?
4) If it doesn't has to then, what should I do in my machine in order to 'enable the port 443. That being said...
5) Should I RUN the AWS Command Line utility from my local machine or from my EC2 remote server?
I'd appreciate some guidance so I can start solving my problem.
Thanks in advance!
You have your region configured as "sa-east-1a", but the region is actually "sa-east-1". Your availability zones are sa-east-1a, sa-east-1b, etc.