* conf/nginx-primaryonly.conf: Use tabs instead of spaces
* management/web_update.py: Includes the tabs so they display with the correct indentation when added to the local.conf
Co-authored-by: 0pis <0pis>
thanks @downtownallday
* this invalidates all user_keys after TOTP status is changed for user
* after changing TOTP state, a login is required
* due to the forced login, we can't and don't need to store the code used for setup in `mru_code`
* Due to the way that the /login UI works, this persists at least one failed login each time a user logs into the admin panel. This in turn triggers fail2ban at some point.
* hmac.compare_digest() expects arguments of type string, make sure we don't pass None
* Currently, this cannot happen but we might not want to store `mru_token` during setup
* this allows implementation of other mfa schemes in the future (webauthn)
* also makes key management easier and enforces one totp credentials per user on db-level
* Only spawn a thread pool when strictly needed
For --check-primary-hostname, the pool is not used.
When exiting, the other processes are left alive and will hang.
* Acquire pools with the 'with' statement
This will make it so that the HSTS header is sent regardless of the request status code (until this point it would only be sent if "the response code equals 200, 201, 206, 301, 302, 303, 307, or 308." - according to thttp://nginx.org/en/docs/http/ngx_http_headers_module.html#add_header)
These subdomains/records are for automatic configuration of mail clients, but if there are no user accounts on a domain, there is no need to publish a DNS record, provision a TLS certificate, or create an nginx server config block.
When the management commands fail, they can print something to the standard error output.
The administrator would never notice, because it wouldn't be send to him with the usual emails.
Fixes#1763
* Create the mta_sts A/AAAA records even if there is no valid TLS certificate because we can't get a TLS certificate if we don't set up the domains.
* Make the policy id in the TXT record stable by using a hash of the policy file so that the DNS record doesn't change every day, which means no nightly notification and also it allows for longer caching by sending MTAs.
Folks didn't want certificates exposing all of the domains hosted by the server (although this can already be found on the internet).
Additionally, if one domain fails (usually because of a misconfiguration), it would be nice if not everything fails. So grouping them helps with that.
Fixes#690.
with this nginx will keep on proxying requests and serve static content
instead of passing this responsibility to proxied server
Without this the one needs to run an additional server to server static
content on the proxied url
removed explicit call of the system python, cause the file has a shebang with the mail-in-a-box shipped python.
for me the system python complaint, that it is missing some modules
Previously the notification email sent when a box's SSL certificate
is automatically updated said, "Error Provisioning TLS Certificate"
even when there was no error. This changes the subject line to "TLS
Certificate Provisioning Results", which is more accurate.
it is sometimes needed to be able to set axfr to more than just one ip address. This can be done with multiple xfr: in the secondary dns input but if you need to add an entire subnet segment (xxx.xxx.xxx.0/yy) then it will not work.
With this patch it is now possible to use a subnet as input for xfr the same way as if it was an ip address.
The resulting command had nested single quotes which doesn't work
I think this fixes all/most of the issues in #1627. I am getting a full backup, then the next time it's run I get an incremental. running from the CLI with --status looks good, --verify looks good, and --list looks good.
The elif needed to check to see if the string was in the listing of results of the shell command. As it was the conditional was just the string which always evaluates to true and was therefore giving a misleading error message.
* drop the ondrej/php PPA since PHP 7.x is available directly from Ubuntu 18.04
* intall PHP 7.2 which is just the "php" package in Ubuntu 18.04
* some package names changed, some unnecessary packages are no longer provided
* update paths
@joshdata squashed pull request #1398, removed some comments, and added these notes:
* The old init.d script for the management daemon is replaced with a systemd service.
* A systemd service configuration is added to configure permissions for munin on startup.
* nginx SSL settings are updated because nginx's options and defaults have changed, and we now enable http2.
* Automatic SSHFP record generation is updated to know that 22 is the default SSH daemon port, since it is no longer explicit in sshd_config.
* The dovecot-lucene package is dropped because the Mail-in-a-Box PPA where we built the package has not been updated for Ubuntu 18.04.
* The stock postgrey package is installed instead of the one from our PPA (which we no longer support), which loses the automatic whitelisting of DNSWL.org-whitelisted senders.
* Drop memcached and the status check for memcached, which we used to use with ownCloud long ago but are no longer installing.
* Other minor changes.
Passwords must be eight characters long; when passwords are changed via the users page the dialog states that passwords need to be at least four characters but only eight or more are acceptable.
The cryptography package has created all sorts of installation trouble over the last few years, probably because of mismatches between OS-installed packages and pip-installed packages. Using a virtualenv for all Python packages used by the management daemon should make sure everything is consistent.
See #1298, see #1264.
* The Mozilla recommendations must have been updated in the last few years.
* The HSTS header must have >=6 months to get an A+ at ssllabs.com/ssltest.
* [Issue #1159] Remove any +tag name in email alias before checking privileges
* Move priprivileged email check after the conversion to unicode so only IDNA serves as input
* Fixed issue with relative path for rsync relative names
Actually using the parsed URL `path` part, instead of doing a lousy split().
Renamed the `p` variable into something more sensible (`target`).
Fixes: #1019
* Added more verbose error messages upon rsync failures
fixes#1033
* Added command to test file listing
* Added support for backup to a remote server using rsync
* updated web interface to get data from user
* added way to list files from server
It’s not using the “username” field of the yaml configuration
file to minimise the amount of patches needed. So the username
is actually sorted within the rsync URL.
Signed-off-by: Bernard `Guyzmo` Pratz <guyzmo+github@m0g.net>
* Added ssh key generation upon installation for root user.
Signed-off-by: Bernard `Guyzmo` Pratz <guyzmo+github@m0g.net>
* Removed stale blank lines, and fixed typo
Signed-off-by: Bernard `Guyzmo` Pratz <guyzmo+github@m0g.net>
* fix backup-location lines, by switching it from id to class
* Various web UI fixes
- fixed user field being shadowed ;
- fixed settings reading comparaison ;
- fixed forgotten min-age field.
Signed-off-by: Bernard `Guyzmo` Pratz <guyzmo+github@m0g.net>
* Added SSH Public Key shown on the web interface UI
Signed-off-by: Bernard `Guyzmo` Pratz <guyzmo+github@m0g.net>
* trailing spaces.
Signed-off-by: Bernard `Guyzmo` Pratz <guyzmo+github@m0g.net>
* fixed the extraneous environment
Signed-off-by: Bernard `Guyzmo` Pratz <guyzmo+github@m0g.net>
* Updated key setup
- made key lower in bits, but stronger (using -a option),
- made ssh-keygen run in background using nohup,
- added independent key file, as id_rsa_miab,
- added ssh-options to all duplicity calls to use the id_rsa_miab keyfile,
- changed path to the public key display
Signed-off-by: Bernard `Guyzmo` Pratz <guyzmo+github@m0g.net>
* added rsync options for ssh identity support
Signed-off-by: Bernard `Guyzmo` Pratz <guyzmo+github@m0g.net>
* removed strict host checking for all backup operations
Signed-off-by: Bernard `Guyzmo` Pratz <guyzmo+github@m0g.net>
* Remove nohup from ssh-keygen so errors aren't hidden. Also only generate a key if none exists yet
* Add trailing slash when checking a remote backup. Also check if we actually can read the remote size
* Factorisation of the repeated rsync/ssh options
cf https://github.com/mail-in-a-box/mailinabox/pull/678#discussion_r81478919
* Updated message SSH key creation
https://github.com/mail-in-a-box/mailinabox/pull/678#discussion_r81478886
* sshfp records from nonstandard ports
If port 22 is not open, dns_update.py will not create SSHFP records
because it only scans port 22 for keys. This commit modifies
dns_update.py to parse the sshd_config file for open ports, and
then obtains keys from one of them (even if port 22 is not open).
* modified test of s per JoshData request
* edit CHANGELOG per JoshData
* fix typo
Seems like if REQUEST_METHOD is set to GET, then we can drop two redundant ways the query string is given. munin-cgi-graph itself reads the environment variables only, but its calls to Perl's CGI::param will look at the command line if REQUEST_METHOD is not used, otherwise it uses environment variables like CGI used to work.
Since this is all behind admin auth anyway, there isn't a public vulnerability. #914 was opened without comment which lead me to notice the redundancy and worry about a vulnerability, before I realized this is admin-only anyway.
DavDroid's latest version's account configuration no longer just asked for a hostname. Its email address & password configuration mode did not work without a SRV record.
adding :
target="_blank"
to
<li><a href="/admin/munin">Munin Monitoring</a></li> on line 96
Why ?
Because when you click on munin link, and follow links, you lose your index, or click back many times...
So i propose my pull request.
Et voilà ^^
On some machines localhost is defined as something other than 127.0.0.1, and if we mix "127.0.0.1" and "localhost" then some connections won't be to to the address a service is actually running on.
This was the case with DKIM: It was running on "localhost" but Postfix was connecting to it at 127.0.0.1. (https://discourse.mailinabox.email/t/opendkim-is-not-running-port-8891/1188/12.)
I suppose "localhost" could be an alias to an IPv6 address? We don't really want local services binding on IPv6, so use "127.0.0.1" to be explicit and don't use "localhost" to be sure we get an IPv4 address.
Fixes#797
This was originally fixed in 143bbf37f4 (February 16, 2015). Then I broke it in 7a93d219ef (November 2015) while doing some refactoring ahead of v0.15.
https://discourse.mailinabox.email/t/status-check-emails-empty-after-upgrading-to-v0-16/1082/3
A user on that thread suggests an alternate solution, adding `PYTHONIOENCODING=utf-8` to `/etc/environment`. Python docs say that affects stdin/out/err. But we also use these environment variables elsewhere to ensure that config files we read/write are opened with UTF8 too. Maybe all that can be simplified too.
0pis
Joshua Tauberer
Felix Spöttel
David Duque
Richard Willis
Vasek Sraier
A. Schippers
Michael Becker
Sumit
Jarek Jurasz
Matthias Hähnel
Edwin Schaap
Dan Jensen
notEvil
Victor
Kim Schulz
Snacho
captainwasabi
jvolkenant
Michael Heuberger
Pascal Garber
mbraem
Ryan Stubbs
jeff-h
Christopher A. DeFlumeri
hlxnd
NatCC
John Olten
Jānis (Yannis)
Nicolas North
Fabian Bucher
Marius Blüm
Git Repository
Michael Kroes
François Deppierraz
Rinze de Laat
Théo Segonds
Sean Watson
Ian Beringer
Dominik Murzynowski
Jonathan Chun
guyzmo
rxcomm
Steve Gregg
Mathis Hoffmann
Scott Bronson
cs@twoflower
schlypel
Rinze
Arnaud
aspdye
Marc Schiller
msgerbs
Robert G. Jakabosky
Wolf-Bastian Pöttner
mike
Jeroen Jacobs