I love Dgraph so much! Thank you all on the front lines!
May I get some help deploying to a Vultr Cloud please?!
Terminal 1) I SSH in and run
curl https://get.dgraph.io -sSf | bash
dgraph
dgraph zero --my=localhost:5080
Terminal 2) SSH in another terminal and run
dgraph alpha --my=localhost:7080 --lru_mb=1024 --zero=localhost:5080
Terminal 3) Locally run
dgraph-ratel
Terminal 1 Output)
Connected: id:1 addr:"localhost:7080"
Vultr shows us IP information
The firewall I have for the server allows me to SSH in & I opened all the dgraph ports just cause I thought maybe it’d help
& when I’m in Ratel and try connecting to port 8080 all the /health endpoints fail
How may I get to a successful /health please?!
Thank you friends!
Are you able to see the endpoint youADRR:8080/health
? It should be public in that case. Otherwise, you should do some king of proxying with your machine and the cloud.
you should see something like
1 Like
Paras
(Paras Shah)
May 7, 2020, 7:33pm
3
It will be helpful if you can provide a wireshark trace between Ratel and the Alpha on the machine that runs Ratel. This will help narrow it down.
Other things you can try:
A ping test to the Alpha Server IP
A telnet test to the Alpha Server IP:8080
I see that you have opened ports 5080, 7080, 8080, 9080 and 8000. But I am thinking it is some firewall / networking issue at this time but need data.
2 Likes
joaquin
(Joaquin Menchaca)
May 8, 2020, 11:37pm
4
For quick troubleshooting:
Works from localhost?
System Firewall Blocking?
Cloud FireWall Blocking?
1 localhost
ssh into the system, do sanity check:
curl localhost:8080/health
2 local system firewall
Likely this is turned off, but just sanity check. If you have two systems on same subnet, could you try to curl to the local IP address?
Check the distro, e.g.
lsb_release -a
cat /etc/os-release
If you have Ubuntu 18.04, make local firewall is inactive:
sudo ufw status # should say inactive
3 Cloud Firewall
It looks like your cloud firewall config from the screenshot is configured to allow 8080
and 8000
on that system. If the above steps work out as expected, then we can try
curl <public_ip_address_of_vultr_instance>:8080/health
1 Like
Thank you @MichelDiz @Paras & @joaquin ! I really appreciate the support of this mastermind group! I know we’ll find a solution!
I SSH’d in, ran dgraph zero
+ dgraph alpha
& tried
curl localhost:8080/health
[
{
"instance":"alpha",
"address":"localhost:7080",
"status":"healthy",
"group":"1",
"version":"v20.03.1",
"uptime":28
}
]
Plus
cat /etc/os-release
NAME="CentOS Linux"
VERSION="8 (Core)"
ID="centos"
ID_LIKE="rhel fedora"
VERSION_ID="8"
PLATFORM_ID="platform:el8"
PRETTY_NAME="CentOS Linux 8 (Core)"
ANSI_COLOR="0;31"
CPE_NAME="cpe:/o:centos:centos:8"
HOME_URL="https://www.centos.org/"
BUG_REPORT_URL="https://bugs.centos.org/"
CENTOS_MANTISBT_PROJECT="CentOS-8"
CENTOS_MANTISBT_PROJECT_VERSION="8"
REDHAT_SUPPORT_PRODUCT="centos"
REDHAT_SUPPORT_PRODUCT_VERSION="8"
Wireshark!
While not SSH’d
curl vultr.server.ip.address:8080/health
curl: (7) Failed to connect to vultr.server.ip.address port 8080: Connection refused
Chris - Thanks for your nice feedback. We’re thrilled that you like Dgraph. We look forward to working with you.
1 Like
Paras
(Paras Shah)
May 11, 2020, 8:20pm
7
chris-carrington:
Wireshark!
Looking at the wireshark trace, it is clearly a networking issue. Your ICMP Pings are not going through to the ip-address.
Did you start dgraph alpha with the --my option. This is the option you need to provide to bind alpha to the ip:port.
chris-carrington:
–my=localhost:7080
I see this in your earlier comment. Replace this with the vultr-server-ip (not localhost).
1 Like
Thank you for your assistance @Paras ! & thank you for introducing me to Wireshark! Here are the combinations + Wireshark output I’ve done so far
dgraph zero --my=localhost:5080
dgraph alpha --my=vultr.server.ip.address:7080 --lru_mb=1024 --zero=localhost:5080
dgraph zero --my=vultr.server.ip.address:5080
dgraph alpha --my=vultr.server.ip.address:7080 --lru_mb=1024 --zero=vultr.server.ip.address:5080
dgraph zero --my=vultr.server.ip.address:5080
dgraph alpha --my=vultr.server.ip.address:7080 --lru_mb=1024 --zero=localhost:5080
Paras
(Paras Shah)
May 11, 2020, 9:12pm
9
Thank you.
From wireshark, it says “Communication Administratively Filtered”. Per RFC, This generally means an ACL, Firewall or Router on the path from client to server (alpha) is filtering the communication and sending this code
See
Things to do:
1 . You may want to check the Firewall, ACL settings on your client.
2. Maybe try a different client.
3. Take a tracepath/traceroute.
Hope this helps.
2 Likes
Thank you for all the suggestions @Paras ! After I ran into Firewall + ACL dead ends I switched over to AWS and got a successful /health on the first try! Idk why Vultr didn’t work for me but I appreciate your assistance, I learned a lot from you, thank you!
2 Likes
joaquin
(Joaquin Menchaca)
May 13, 2020, 1:36am
11
Congratulations!!! We are happy for your success.
2 Likes
system
(system)
Closed
June 12, 2020, 1:47am
12
This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.