Open jankogaga opened 6 years ago
Do you use a proxy ? A lot of issues are to do with proxy setups (I 've had a few !) If there are no errors in Rocket, then it is unlikely to be Rocket itself. Please go check your server logs - httpd etc.
Yes, we use Nginx proxy. But, the issue never existed last year, but in newer versions. Regarding logs, all I can find: /var/log/httpd/ssl_error_log /var/log/httpd/error_log They are empty.
Thanks, Dragan
Lots of things have changed since last year, not just Rocket....
You probably ought to go and turn on some nginx logging. That may help you. Start here: https://docs.nginx.com/nginx/admin-guide/monitoring/logging/
Thank you. We have already set Nginx logging, so I will back to you once I have another Rocket freezing.
Just it has been frozen, no records in nginx log.
Anything happening at that time on your server when it freezes? How many users connected? How many instances of rocket.chat?
I can't find any log at that frozen time. We have 12 users in total. We have 1 instance.
I have the same problem. Around 10 users, one instance. If I‘m correct the log says something about mongodb, but I have the log not handy right now.
Do you have message pruning or any thing like that setup?
Do you see CPU spikes in relation to this freezing?
I forgot to mention, Rocket chat is dockerized and uses RAM and CPU from the host server and the resources there are not the issue.
Just noticed, when Rocket chat is freezing Process Janus takes 6% of CPU in comparison when it is working well (1-2%)
Greetings, we had such a problem when there was a shortage of RAM, in this case SWAP began to be used, and its volume was too small initially, and when it ended the rocket was frozen. First, I discovered a pattern that this happens when you enter the administration panel from a mobile device. And then it turned out that this was due to a lack of RAM. After rasherieniya RAM and swap everything became good.
Can people advise what amount of RAM they are talking about please?
I had one with 2GB, with that one Rocket.chat was frozen 1x a week. Now it's running on 4GB, seems fine since the last 10 days.
The situation is exactly the same as in the previous commentator, with a 2 GB rocket stably frozen once a week, with 3 GB and increased to 1 GB swap, everything is OK!
I must stress out that my Rocket running in docker has a lot of available RAM on the host node. However, I don't know how to make Rocket to use it. Does any configuration exist for Rocket in order to determine how much RAM to use?
Do we have any recommendation from Rocket Chat team?
The problem is we really don’t have enough information to go on.
If small amount of memory with mongo and rocket.chat in same system. It could be a memory struggle.
Would be interesting to see stats from Prometheus or system graphed so we could see if a pattern associated with freezes
Before installation of any SW on Rocket Chat, could you please tell me if you know how to configure a memory usage of Rocket Chat? Here is the output of free command (one more time Rocket chat is installed as a Docker and it uses all host's resources - there is no HW resource isolation) [root@rocket /]# free -h total used free shared buff/cache available Mem: 62G 17G 376M 101M 44G 44G Swap: 63G 9.3G 54G
Thanks, Dragan
Hi, Here is a graph from Munin monitoring SW, taken at the time of freezing. Please take a look at rocket. The other docker containers are covered.
Is anyone still getting this? If so if you open a new incognito tab.. does the issue persist? Are you able to login from that new tab?
Hi geekgonecrazy,
I am able to log in from an incognito tab. However, freezing doesn't persist just on one session, but for all logged users.
Hi 👋 I get the same issu with rocket behind proxy (Apache2). We're not with a docker instance, it's legacy installation (manual) We only have one instance. We don't have too many users.
We get noting in the logs :
It freeze for all users... send and receipt for clients. and some user get "houston we have a problem..." . Web page is very slow to show up a this moment.
It freeze frome 1-2 minutes. But multiple time in 10-30 minutes...
At the moment when it freeze rocket chat seem to have a log hole in log :
EDIT : They get sent at 13h36 so after 36 minutes waiting ...
After moving proxy from apache to nginx no more freeze is happening. Maybe a websocket timeout... Don't really know.
Hi,
Rocket chat from time to time is frozen. The issue occurs among many versions (the latest I have tried 0.71.1). Rocket chat is running in docker container with Centos 7. I am not able to find anything useful in docker logs. When doing curl, I am able to find the content of a login page.
Thanks, Dragan