Something went wrong, sorry [unresolved, closed]

I have migrated my MiaB install from one provider to another … and most, seemingly, has gone well.

I cannot however view the TLS (SSL) Certificates page in the admin area. I get this error message 100% of the time.

My nginx error log shows the following:

2018/04/13 16:21:21 [error] 24820#0: *186 upstream timed out (110: Connection timed out) while reading response header from upstream, client: 31.42.46.158, server: box.example.net, request: “GET /admin/ssl/status?=1523625620722 HTTP/1.1", upstream: "http://127.0.0.1:10222/ssl/status?=1523625620722”, host: “box.example.net”, referrer: “https://box.example.net/admin

I have attempted to change the timeout in the nginx conf file, but that did not help.

Other than this issue, all is functioning properly.

Any ideas …

Weird, you re-run setup since the migration?

Yeah … I had to delete the ssl directory as the key seemingly did not come through with the backup — as there have been other reports of. I ran the script to issue the certs from the command line so that the system was up with SSL. I am not 100% certain at what point I reran setup but I know that I have at least once. It doesn’t hurt anything to rerun it so I shall try it again in a bit.

Re-ran set up. No change. It did remind me of the Nextcloud issues that are seemingly caused when migrating…

Nextcloud or one of the apps require upgrade - only a limited number of commands are available
You may use your browser or the occ upgrade command to do the upgrade
Set log level to debug
Exception: Updates between multiple major versions and downgrades are unsupported.
Update failed
Maintenance mode is kept active
Reset log level
Nextcloud or one of the apps require upgrade - only a limited number of commands are available
You may use your browser or the occ upgrade command to do the upgrade
Set log level to debug
Exception: Updates between multiple major versions and downgrades are unsupported.
Update failed
Maintenance mode is kept active
Reset log level

But that is another issue … one of no real importance to me as NextCloud is not actively used.

What’s the error output in miab daemon? (systemctl status mailinabox) or /var/log/syslog | grep mailinabox

As ya know from earlier interactions, I am not overly familiar with the various logs … that said …

/var/log/syslog shows nothing relevant … it does show the daily cron jobs for /management/daily_tasks.py and one successful dns entry (see below) … as far as ‘systemctl status mailinabox’, I assume that was a command line that you wanted me to run … and if so, it came back with ‘systemctl: command not found’.

syslog:>

Apr 10 03:00:01 box CRON[19271]: (root) CMD ((cd /root/mailinabox && management/daily_tasks.sh))
Apr 11 03:00:01 box CRON[31202]: (root) CMD ((cd /root/mailinabox && management/daily_tasks.sh))
Apr 12 03:00:01 box CRON[8590]: (root) CMD ((cd /root/mailinabox && management/daily_tasks.sh))
Apr 13 03:00:01 box CRON[28210]: (root) CMD ((cd /root/mailinabox && management/daily_tasks.sh))
Apr 13 23:02:13 box named[15173]: success resolving ‘mailinabox.email/AAAA’ (in ‘mailinabox.email’?) after reducing the advertised EDNS UDP packet size to 512 octets

No luck … I removed the ssl directory and ran the ssl_certificates.py and it failed as it could not find the ssl directory … (odd as I thought that py file was to issue the certs and create the directory). I then reran mailinabox and then NO admin pages would come up! :frowning: After that, I ran the ssl_certificates.py and it issued the cert(s) for the domains which are actually pointed to the box. So now, seemingly back to where I started … but no! The system status checks page did not come up still until I tried all pages, then tried it again and it worked … but still no TLS page.

No don’t remove the directory, remove the contents lol.

Heh, I left that out in my narrative above. Once I tried it, I recreated the directory before rerunning mailinabox. I forgot to mention it though.

1 Like

Since I never posted an update, here it is…

After several tests and reviewing log files looking for an error that we could not find it seems that the box times out if there are over about 20-22 subdomains. It has not been clear exactly what causes the issue but once I cleaned out some unused subdomains things magically began to work normally.

A thank you to @murgero for all of his help in PM.

My conclusion is that although there are no stated specific limits as to how many domains MiaB can handle. There IS a limiting factor somewhere. As this project is not specifically geared towards “enterprise” use and the typical user will have one maybe two domains, I do not know that this issue is worth looking in to further. Maybe @JoshData should decide that.

Apparently, instead, the issue is a change in the way LetEncrypt works – hopefully that can be addressed soon.

@hzink - this issue was unrelated to yours.