@p44 That's very strange. It should definitely save to your Sent folder. Did you perhaps have it set to save sent messages in a different folder? On iOS, you can go to Settings > Mail > Accounts > {accountName} > Account > Advanced and in there you can set the Mailbox Behaviours for Drafts, Send, Deleted, and Archive. Make sure it's set to the correct folder. Sometimes there's a difference between "Sent" and "Sent Messages". There are some semantic differences, though usually I find that only affects Junk/Spam folders as there's been previous discussions on those. The rest of them work fine for me but I may have set it up accordingly too back in the day.

The instructions above assume you're using the built-in Apple Mail app.

@brutalbirdie It was a discount coupon that we sent out via email to some users. It was meant to offer an off on the monthly plan. I changed the title since we didn't send it to most people

@robi said in Cloudron with Nextcloud Talk can't connect calls after v20 upgrade:

We have tested with another Cloudron NC:T (VPS, no NAT) and the same user from the same NAT network, and we could connect a call.

This seems then something TURN and NAT related. I think non firefox, you can go to about:webrtc to see what is happenning

Will mark this as solved since this is a DO issue.

@girish Thanks, after looking through a long list, turns out it was Snipe-IT, that we don't use and I just had installed as a test. So stopping it for now but maybe worth knowing in-case the pp itself is a but hungrier than it should be for some reason.

@vladimir I can also recommend sending a test mail to https://mail-tester.com to check if you have setup the DNS records required for the external email server correctly.

@d19dotca said in Outbound email error: "Client network socket disconnected before secure TLS connection was established":

Are you aware of any method for me to determine what the receiving mail server supports for TLS protocols and cipher suites?

You can try something like openssl s_client -starttls smtp -connect server:25. Note that usually you have to do this from a real server and not from your laptop, since most residential networks will block outbound port 25.

There is also a better tool for SMTP connection testing, it's at the tip of my tongue...

Edit: gnutls-cli is the tool. It has some feature which openssl s_client does not support. I can't seem to remember what exactly.

@mehdi you're right, my poor choice of words..

apt is a binary that can call libapt-pkg libraries and provides a subset of interactive options and commands for apt-get/cache by default.

apt-get is not limited to those options.

@robi said in Firewall IP blocking: IPv6 not possible:

Marcel, can you share more detail about your chosen block list and how others can do the same?

Sure!

top 10 countries of attacks: https://www.privacyaffairs.com/geopolitical-attacks/

Source of country ip's: https://www.ipdeny.com/ipblocks/

I've choosen to only block those below, we don't expect any necessary traffic from those countries (it's more than 45% of the known Countries where attacks come from):

China: https://www.ipdeny.com/ipblocks/data/aggregated/cn-aggregated.zone

Russia: https://www.ipdeny.com/ipblocks/data/aggregated/ru-aggregated.zone

North Korea: https://www.ipdeny.com/ipblocks/data/aggregated/kp-aggregated.zone

Iran: https://www.ipdeny.com/ipblocks/data/aggregated/ir-aggregated.zone

Pakistan: https://www.ipdeny.com/ipblocks/data/aggregated/pk-aggregated.zone

Syria: https://www.ipdeny.com/ipblocks/data/aggregated/sy-aggregated.zone

India: https://www.ipdeny.com/ipblocks/data/aggregated/in-aggregated.zone

Vietnam: https://www.ipdeny.com/ipblocks/data/aggregated/vn-aggregated.zone

All those IP's copy-pasted in Cloudron > Network> Firewall, currently 20906 ranges blocked.

I added them this morning and I can tell now already that spam has reduced with 90%

@nicolas There's a feature request to monitor network traffic at the app level. If we had that, we could have narrowed this down more quickly.

This is fixed in Cloudron 6

@yusf Are backups failing by any chance. I wonder why there are multiple backups in a day. How often do you create backups? If it's one week backups from 17th Npv are expected. the ones which are slightly older are probably preserved for various reasons. One thing you can do is Backups -> Cleanup backup. Then in the same view, click on logs and it will tell you why each and every backup is preserved. If you paste the log, I can explain in a bit more detail.

@dreamcatch22 with AWS, I have usually found that it's either

a) the cpu credits are kicking in. are you aware of https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/burstable-credits-baseline-concepts.html ? AFAIK, even lightsail has the same limitation (but not 100% sure).

b) some machine have a lot of memory but perform very poorly . For example, see the notes in EC2 R5 xlarge (which has 32GB RAM!) but works very poorly with docker - https://forum.cloudron.io/post/17488

Not sure, if either of those apply to you. What instance are you using ?

@nebulon Hi, I have Cloudron running in a VM under Proxmox, and since I (rather poor decision on my part) didn't take a backup of the VM before installing Cloudron, I don't easily have a "restore point". Whenever I install a new VM image, I have to manually configure the networking to use the right bridge, so just a hassle. Was hoping to avoid it, but clearly it's unavoidable.

But understood, I ended up just deleting the VM's disk image and remaking it, just took awhile.

tldr - my Cloudron install is on a VM on a dedicated server, not a VPS, and no restore point, so can't easily just "rollback".

@girish this works!!! Thanks for the hint.

This is actually intentional as the system waits for the healthcheck.

Support for Ubuntu 20.04 will come with next release.

@robi Whoops indeed, I was lost in between terminals :))

@d19dotca Yes, the limits are there to protect against the noisy neighbor problem which exists when many processes are competing for the same resources and ONE uses up more than their fair share.

Technically we could have all 30 Apps be set to 1+GB on a 16GB RAM system and it would work fine until one App behaved badly. Then the system would be in trouble as the OOM killer would select a potentially critical service to kill.

With limits, the system is happy, and the killing happens in containers instead.