New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
SolusVM master is losing connection with the slave
A few days ago I had to migrate the master from one server to another in different DC, the migration was ok, database was restored with no problem, slaves were up and running non stop, now after the migration the master is losing connection with one slave only and send me warning emails, I had my DCs run mtr test between them and I did also myself between the IPs and we found no dropped packets, SolusVM support gave instruction to disable the warnings, but did not work still receiving some warnings emails, it reduced a little but did not stop.
The thing is I have more than one slave on that DC and the others slaves were not affected.
I would like to know if you guys had experienced something like that.
Comments
Check your email header. ensure that it is not sending out from your old master just to be sure.
What's CPU load, RAM use and network traffic like on the node? It might be worth monitoring this, SolusVM gets a bit crazy when your node does.
Did you run an mtr from both sides?
Was happening to me a while back. Cause was some kids were ddosing the server and it saturated the network port.
Try a ping6 between them. Sometimes the master will get ipv6 for the slave's domain resolvement and if you didn't config it yet, it will take a long time response and thus the warning.
Thank you, the node are new has only 17% load, the cpu load is very low and memory is about 14% used.
DC did mtr both sides and I did myself on the IPs and no packet lost on both cases.
Thank you, the node is new, has few VPSs there and we did not detect any DDoS attack, this starts to happens when I migrate the master to another server, also all VPSs are up and running they are not dropping only the communication between this node and master, weird that I have another nodes on the same DC connecting to the same master and they are ok.
thank you, I will do it
Just receive emails from the host and look at the header and they are from the new master IP
you need to update your new slave's master IP, was that done? /usr/local/solusvm/data/allow.dat
how bout licensed IP?
Thank you, it was done
If the slave is OpenVZ based it could be a hung container which creates an unkillable vzctl process within a container, the only fix is to reboot the node, you can pretty much confirm this is the issue by restarting lighttpd on the slave or fpm if you upgraded to nginx, if the problem goes away for 30 minutes or so that is probably the issue.
Often the problem is finding out which container was the cause because if you don't find that out first you will end up in the same situation after reboot if the same conditions occur.
Can you reply to your support ticket and ask for Phill to take a look please.
Thank you, sure do
Phill is awesome!
Hi General, did already waiting for him to answer, thanks
Let me know how it goes
SolusVM support is awesome. They will resolve your issue in no time.
You are right, they are, always a help hand there
They assign me to a L3 support and they change something on the slave that I don't know what they did yet and asked me to follow if the messages stop, I will do it and if everything is ok I will ask them what they did so I can post and can help anybody else that would experience the same issue
Some update, after a these days finally they found the issue, I want to tanks Harsha from SolusVM, for his time and for not giving up, great guys.