Runners are best run outside of Cloudron. Just get a separate VPS and run it there. This is the correct approach for security and performance . For example, git.cloudron.io is on a Cloudron. There is a gitlab runner for all the package building and running tests. It is on a separate hetzner dedi.
Our runner is the docker runner. Cannot imagine giving the runner access to docker which is having other production apps.
@mathieu you can clone this repo https://git.cloudron.io/packages/calcom-app , apply the changes in the Dockerfile as needed and build the package yourself.
@jcalag I think I have replied to your email.
@difficult222 I have replied to your ticket.
@dd4y if the instructions I gave in the previous post worked, we can adjust the package to use a larger heap size by default (heap size is adjusted using max-old-space)
Yes, you have to adjust the schedule manually to not overlap. Otherwise, if a single app is backing up to multiple sites in parallel, they will get serialized internally (with the lock).
64GB seems like a ridiculous amount of space for a web app!
@dd4y Can you this:
export NODE_OPTIONS="--max-old-space-size=4096"I think that value can go up to 16GB . Can you try by increasing that value till it works ?
Usually, that error happens when docker became corrupt (for example, server ran out of disk space). Is that a possibility in your case? This is quite an old version of Cloudron. The recent versions of cloudron-support have --recreate-docker but I don't think it's there ins Cloudron 7 . If you write to support@cloudron.io , I can look into it.
@timconsidine thanks for updating. On the pricing page, yeah, it says just 1 user allowed. Otherwise, 10/user/month.
@timconsidine does the pricing at https://appflowy.com/pricing mean that it is 10/user/month for selfhosted?
@difficult222 there is something wrong with that output . It has 2026-01-24-060000-602 and snapshot . These directories should not be there (in fact, they are part of backups).
Can you check if those directories exist in the Hezner S3 also? If it did, something has gone wrong with how the backups are getting created . Could it be that you changed the prefix of a site? Or used a prefix that already had some contents from previous backup setup?
@kubernetes also --reenable-ipv6
@kubernetes if it helps, the internet network is documented here - https://docs.cloudron.io/network#internal-network
/var/log/unattended-upgrades/unattended-upgrades.log has the ubuntu unattended upgades log
Does Docuseal have a button to test email sending?
gelocation also works. But I am surprised that it blocks login. Let me check this internally.
$ ping geolocation.cloudron.io
PING geolocation.cloudron.io (167.99.137.196) 56(84) bytes of data.
64 bytes from my.public.cloudron.io (167.99.137.196): icmp_seq=1 ttl=56 time=18.3 ms
64 bytes from my.public.cloudron.io (167.99.137.196): icmp_seq=2 ttl=56 time=31.3 ms
64 bytes from my.public.cloudron.io (167.99.137.196): icmp_seq=3 ttl=56 time=17.2 ms
165.227.67.76 is cloudron.io 's IP . It looks like the server cannot connect to cloudron.io ? Can you ping 165.227.67.76 on the server?
$ ping cloudron.io
PING cloudron.io (165.227.67.76) 56(84) bytes of data.
64 bytes from prod.cloudron.io (165.227.67.76): icmp_seq=1 ttl=53 time=114 ms
64 bytes from prod.cloudron.io (165.227.67.76): icmp_seq=2 ttl=53 time=127 ms
64 bytes from prod.cloudron.io (165.227.67.76): icmp_seq=3 ttl=53 time=310 ms
@charlesnw can you try increasing the memory limit for that redis? Services -> redis ->memory limit?
@Supaiku said in Graphite & SFTPwon't run?:
I'm honestly not quite clear what graphite does, so I'm curious about that.
Graphite helps with the graph rendering to collect stats.
I don't use SFTP much, but still want thing runinng smoothly and didn't exactly figure out what the initial issue actually was...
SFTP provides file browsing services for the containers i.e the Filemanager.
Are the containers in "recovery mode" ? Both the containers are dying gracefully, so someone is stopping them explicitly ("waiting for xxx to die).
@pauliusl Debug this step by step: host my.domain.com on your laptop. Does it point to your server IP?
ERR_NAME_NOT_RESOLVED
When you are moving domains across provides this is quite common. The whole migration process is not atomic.