d00p Posted April 10, 2019 Posted April 10, 2019 Dear Froxlor Commuity, finally - the first release candidate of our new API based version 0.10.0! A lot of work has gone into this, many internal changes (you might miss any frontend-changes, but be patient...) most importantly the API backend which not only is used by froxlor frontend itself but can also be uses from within your website/scripts/etc. Froxlor now uses composer to include some of its requirements like phpMailer, Logger, IdnaConvert and TwoFactorAuth libraries. Here are some of the new features besides API that found their way in: - 2FA / TwoFactor Authentication for accounts - MySQL8 compatibility - new implementation of Let's Encrypt (acme.sh) - customizable error/access log handling for webserver (format, level, pipe-to-script, etc.) - lots and lots of bugfixes and small enhancements You can see all changes on Github at https://github.com/Froxlor/Froxlor/compare/0.9.40.1...0.10.0-rc2 Download: 0.10.0-rc2Note: There will be no Debian packages for release-candidates.Visit http://www.froxlor.org or join our IRC channel #froxlor on irc.freenode.net.Thank you,d00p
Gamerboy59 Posted May 1, 2019 Posted May 1, 2019 That's amazing, now we need an api doc to work with it. Great work!
d00p Posted May 1, 2019 Author Posted May 1, 2019 16 minutes ago, Gamerboy59 said: That's amazing, now we need an api doc to work with it. Great work! https://api.froxlor.org/
crazy4chrissi Posted May 16, 2019 Posted May 16, 2019 That sounds and looks great. Good work! I think in the longrun it would make sense to separate the Froxlor frontend from the backend. I guess it could make sense to install the froxlor backend without the frontend, but use your own frontend. Or someone may come of with a CLI tool that can talk to the froxlor API and you don't need any webfrontend at all. That would be really cool. As I guess there are no other backends that support the same API, it does not really make sense to use the froxlor frontend without the backend. So I see the frontend as a component that requires the backend, but the backend as a component that could be used standalone.
d00p Posted May 16, 2019 Author Posted May 16, 2019 As a matter of fact, there is no dependency from the API to the front-end really...all the sanity checks etc. Are in the API, not the front-end
d00p Posted June 21, 2019 Author Posted June 21, 2019 As there is little to no feedback for this release candidate I cannot tell for sure as I need to test as much as possible myself
cscholz Posted June 22, 2019 Posted June 22, 2019 (edited) I've update my prod buster system today to 10.0-rc2. I'll report any issues i'll find. How can I report possibl bugs? Edited June 22, 2019 by cscholz
d00p Posted June 24, 2019 Author Posted June 24, 2019 On 6/22/2019 at 8:16 PM, cscholz said: How can I report possibl bugs? Either https://forum.froxlor.org/index.php?/forum/15-0100-api/ (if questions or unsure of usage) or https://github.com/Froxlor/Froxlor/issues if you are certain it's a bug
snoopotic Posted July 7, 2019 Posted July 7, 2019 Is there something special to be mentioned at upgrade from 0.9.40.1 that is not stated here: https://github.com/Froxlor/Froxlor/wiki/Updating-Froxlor ? do the cronjobs stay and work etc.?
d00p Posted July 7, 2019 Author Posted July 7, 2019 Yes, this stays all the same. Just update as usual
snoopotic Posted July 7, 2019 Posted July 7, 2019 Yeah. Just did it. First view: looks good, good job! One thing I like to mention: In my Environment I use a lot of LE-Certs. On Upgrade the ALL get renewed. By default it's a 5 minute cron for each webserver-change and each LE-renewal run... This means: after Upgrade all certs get deleted and thus web server configfiles recreated with HOSTS default certs. within this webserver-recreation all LE certs get recreated. If they are not fully recreated within 5 minutes the following job: webserver-renewal with the LE-cert paths may take another 5 minutes. For me: My sites were 15 minutes without invalid certs (not that ugly as I have not that highly visited pages :P) Maybe we can add a hint that it may mean up to a 10-15 minutes "possible outage" on upgrade!?
d00p Posted July 7, 2019 Author Posted July 7, 2019 This is due to the change of the acme implementation. I see what I can do there
Tealk Posted July 12, 2019 Posted July 12, 2019 Are there any points I can test? On 7/7/2019 at 3:20 PM, snoopotic said: Maybe we can add a hint that it may mean up to a 10-15 minutes "possible outage" on upgrade!? After that, your pages went back?
d00p Posted July 12, 2019 Author Posted July 12, 2019 7 minutes ago, Tealk said: Are there any points I can test? Installation, update of 0.9.x to 0.10.0-rc or better current git, adding, editing and deleting of entities, paying around with external API call, etc.etc.etc.
Tealk Posted July 12, 2019 Posted July 12, 2019 So there were no problems with the update (Debian 9) I could also create and edit subdomains as well as delete them. 2FA also works I also installed it on a test server without any problems. (Debian 10) On the server I'm still testing
snoopotic Posted July 14, 2019 Posted July 14, 2019 @Tealk yup, after that pages went back with new valid certs. Well, maybe I found a bug but I will search for corresponding logs but maybe reporting the use case is quicker: - Create E-Mail Address -> ok - Create E-Mail Account -> fails with: Fehlermeldung Die E-Mail-Adresse "" enthält ungültige Zeichen oder ist nicht vollständig. looks like a typo as the address is already "known" ... shall I look further or is this possibly a case? (I just upgraded my Debian 9 to 10 maybe it's a cross over thingy but I don't think so other things work) EDIT: Oh Noes. Horses back: Entering an alternate E-Mail is mandatory xD
d00p Posted July 14, 2019 Author Posted July 14, 2019 34 minutes ago, snoopotic said: EDIT: Oh Noes. Horses back: Entering an alternate E-Mail is mandatory xD It is?? Just checked, it's not: * @param string $alternative_email * optional email address to send account information to, default is the account that is being created And I cannot reproduce this on my Testsystem. Can create account without any alternative address. Please provide detailed information and steps to reproduce
Tealk Posted July 29, 2019 Posted July 29, 2019 I just upgraded my main system to buster and an error occurred with "bind". In the "named.conf.local" the include of Froxlor was included twice. // // Do any local configuration here // // Consider adding the 1918 zones here, if they are not used in your // organization //include "/etc/bind/zones.rfc1918"; include "/etc/bind/froxlor_bind.conf"; include "/etc/bind/froxlor_bind.conf";
d00p Posted July 29, 2019 Author Posted July 29, 2019 Did you run the configuration steps twice? Afaik it just adds the last line...
d00p Posted July 29, 2019 Author Posted July 29, 2019 Well the config template does it only once: https://github.com/Froxlor/Froxlor/blob/master/lib/configfiles/buster.xml#L366 So maybe it was already in there before your upgrade to buster and you re-ran the config steps or similar
d00p Posted July 29, 2019 Author Posted July 29, 2019 No initial configuration of services is done automatically
Recommended Posts
Create an account or sign in to comment
You need to be a member in order to leave a comment
Create an account
Sign up for a new account in our community. It's easy!
Register a new accountSign in
Already have an account? Sign in here.
Sign In Now