hi, and i appreciate and welcome any feedback! Our discourse instance crashes every once in a while and we can not figure out what is causing it. Our discourse is updated to the latest version and just for the record it was created with the one click install from digital ocean running on 4gigs of RAM. please help. here is a shot of the logs . This is urgent so please ask questions if you need to and i will provide as much info as i can! Thanks so much!
Make sure you are on the very latest version.
Does your droplet have a swapfile?
can you explain what you mean by “crashing”? What happens? What does memory on the box look like in that state? What plugins do you have running?
@codinghorror yes we are upgraded to the latest release: v1.6.0.beta11 +58
We experienced this about a month ago, and we “solved” it by moving from a 2GB RAM Digital Ocean box to a 4GB one.
Another detail that might be helpful: we originally installed Discourse using Digital Ocean’s one-click install.
Also, we have not made any customizations to Discourse at all outside of tweaking options available in the admin panel.
@ron_jeremy where would a swapfile normally be located? We can chec
Upgrade again to +80, we had some issues late last week with the tests passed branch
If it “white screen” or errors with 500 I need to see logs and memory status in the droplet
In the Linux terminal, run:
Your swapfile will be listed if you have one. You can create one if you don’t see it listed (you will need admin privs):
In the below example, we will be creating a 4GB swapfile file:
sudo fallocate -l 4G /swapfile
Now we enable the file:
sudo chmod 600 /swapfile sudo mkswap /swapfile sudo swapon /swapfile
Now we make the swapfile permanent:
sudo nano /etc/fstab
At the bottom of the file add the following line entry:
/swapfile none swap sw 0 0
If you run free -h again you should see the swapfile
@sam we updated a couple of days ago and it shows v1.6.0.beta11 +58. When we do an update does it not pull you to the latest version? Just want to get as much info as I can for future events.
Also, when it “crashed” yesterday, we knew this by going to our URL and being presented with a white screen. Power cycling the server seem to have fixed it. Would you be okay with me sending you logs to your inbox? Our logs could contain sensitive information.
@ron_jeremy we will check up on the swap to see if it is enabled.
All, Like I said in the beginning, thanks for the great response! You guys are awesome and we truely appreciate it! Happy coding
Here is a screenshot of the most recent logs, I hope this helps. If you notice anything out of the ordinary please let me know. Also, we can provide more information if need be. Thanks again for all your help !
Notice the fatal alerts in this screenshot:
we also have a lot of spoof attack alerts that are critical, i did not upload a screenshot because it contains ip addresses.
When you say “upgrade again to 80+” the admin panel is showing I am already up to date, and when I go into the repo and pull master, there has only been one commit since July 7
commit cad2919f12b414b229b8c3e2a56baa5c3f4e40a3 Author: Guo Xiang Tan <firstname.lastname@example.org> Date: Fri Jul 15 15:22:15 2016 +0800 Add config options to skip assignment of a mac address.
How do I go about manually upgrading to 80+?
Discourse just crashed again.
I restarted and created a swap file as you instructed. Now when I run free -h I see:
root@discourse-chani:/var/discourse# free -h total used free shared buffers cached Mem: 3.9G 3.4G 512M 275M 31M 1.5G -/+ buffers/cache: 1.8G 2.0G Swap: 2.0G 0B 2.0G
Does this look like it’s properly configured? So once the memory fills up it will start using swap, and this will prevent further crashes?
Go to /admin/upgrade and upgrade from there to make sure you are on latest.
Be sure to upgrade this thing from the console at least once
cd /var/discourse git pull ./launcher rebuild app
OK - we already were on latest (I upgraded a few days ago):
Did you upgrade via the web UI, or from the command line?
If it gets into a crashed state in the next 8 hours can you PM me I can log in and have a look.
I usually upgrade via the web UI, but sometimes I will do it from the command line by doing a git pull and restarting. It looks like we are currently only one commit behind master, and next time I restart, that commit will be live as well.
By the way, here are our memory use stats a few hours after a restart. Does this look accurate? Should the system be reclaiming RAM?
Our forum completely crashed again yesterday. We are trying to refrain from touching the files within the container but could this have something to do with unicorn workers ? Do you have any other suggestions as to what we can check to provide you with information to make an educated guess on what is causing the issue? I have searched a lot and can’t seem to dig anything up. Thanks for being patient with me!
We’re continuing to crash about once every 24 hours.
Here are Discourse’s logs from our crash an hour ago:
I’m PM’ing you as well about this.
If you are using plugins that discourse does not ship as default please turn them off and rebuild immediately. That is step ZERO.
basically, the only thing we have done to discourse is, install it on a droplet using Digital Ocean’s “One-Click Application Install”. As for plugins, i know that i have not installed any and i don’t think @ossia has installed any but i will let him speak for himself.
Here is a screen shot of plugins that are installed