Beiträge von ñull

    Further information for this bug.


    Liveconfig's database is under MySQL and when I look in the table HOSTINGCONTRACTS then I can find back a row for this client. So what happens is that in the web frontend the contract disappears from the list but the backend the ccontract still exists. I will change the title accordingly.

    When you create a new FTP account, you can indicate, beside username and password, also the path. When the FTP accounts is already created, there is no way to verify the path because it is not listed, nor visible when you edit the password. Please could you add this small feature? It will help us in user support and it will help the user to know where he will access the server.

    Two use cases:


    • The typical emails automatic notification emails with no-reply@emailaddress.com. Probabably will have a reply-to header, but some user can still start using it and I just don't want to clean out the mailbox
    • A billing software like WHMCS can only associate clients with email addresses. Some clients just don't want to receive emails or don't want to give an email address. I create a unique email for them, because WHMCS cannot do without, but I don't want to see emails nor see them bounce.


    Both make sense to me and while they might be illegal in the strict sense, nobody is really bothered by this use.


    Meanwhile I found a way to create a blackhole address:


    • Add this line to /etc/aliases:

      Code
      null: /dev/null


    • Compile the changes by running:

      Code
      newaliases


    • Make a custom virtual alias file /etc/postfixt/virtual_alias_custom with:

      Code
      null@existingmaildomain.com  null@localhost


    • Take care that the file permission and owner are the same as /etc/postfix/virtual_alias and then run in the /etc/postfix directory:

      Code
      postmap virtual_alias_custom


      virtual_alias_custom.db should appear now.

    • Modify /etc/postfixt/main.cf to include this new file:

      Code
      virtual_alias_maps = hash:/etc/postfix/virtual_alias hash:/etc/postfix/virtual_alias_custom


    • Restart Postfix

    The translation of this is impossible in Spanish and even in English it is shaky, especially when the variable is used. As example translation of the string "Invalid secondary name server #%i selected".


    In English it is linguistically weird to speak about a primary, first secondary, second secondary or third secondary server. It would be better to speak about first, second, third, fourth etc.


    What is weird in English is utterly rediculous in Spanish. In Spanish nobody says first secondary, second secondary. Unless you change things I cannot find a good way to the example string to Spanish and I am sure there are other languages where is not acceptable.

    Because of the previous bug I wanted to password protect my database maintenance tool which is installed as an application. "Password-protected directories" does allow you to enter an absolute path /var/www/user/apps/myapp (I did not see any error message) but this won't protect it.


    Could you please add a feature to also protect application paths?

    I had to temporary disable a client domain. So I set both HTTP and HTTPS to "Webspace disabled". Accessing the HTTP URL it will show correctly the "Domain ... is not available" notice, but HTTPS won't! In stead it shows another website that has HTTPS active. In my case it rather inconveniently shows the database administration tool.


    I know I should rely on "security through obscurity" but still the same I think we should not make it too easy to find this tool! The expected behaviour is that both HTTP and HTTPS show the same not available page.


    This is for Ubunu 12.04 LTS and Apache webserver.

    You can edit the recovery email template in your LiveConfig settings.


    You were right, that is part of the issue, my Spanish translation is not finished yet and therefore I need to override the default as a temporal work around.


    Then still there is another issue. Turns out that the password reset email always follows the account preference language and it ignores the log-in language override.


    For instance there is this paranoid hosting company that never sends password by emails in an attempt to prevent hijacking. The new user never logged in and his user preference is set to English. The user is from Spain and hardly knows English. He switches to Spanish before resetting his password and gets all the messages in his preferred language except the email, because his preference is still set to English.


    For a consistent user experience, please let log-in language override also work for the Password reset email.

    I try to log-in, forgot my password, change to my language (I tried Spanish), click the password recovery link and get nice instructions in my language. Successfully submit the username and captcha and again get a pleasant response in my language. Then the email arrives and, great disappointment, it is in English.


    Correct this please, because people who change language definitively will have trouble reading the undesired language.

    Ich lese im Handbuch:


    Zitat

    Einrichtung „eigener Links“


    Als Administrator (ab späteren LiveConfig-Versionen auch als Wiederverkäufer) können Sie die „eigenen Links“ direkt über die LiveConfig-Oberfläche unter Verwaltung → LiveConfig → Eigene Links konfigurieren. Diese Links werden jeweils allen Benutzern auf der selben Verwaltungsebene sowie allen eigenen Kunden (ersten Grades) angezeigt.


    Frage ist wie man eigener Links (und damit Funktionen) für alle Kunden bereit stellen kann? Das ist anscheinend nicht möglich. Warum nicht? Zum Beispiel wollte ich gerne meine eigene Backup Funktion dort einrichten für alle Kunden (eigenen oder nicht). Ich habe sogar auf Administrator Ebene nur zwei Wiederverkäufer.


    Wann wird das für Wiederverkäufer möglich sein?

    A client has abusive scripts in their hosting account. I don't want to deactivate th whole domain, because they still have right to use their email (they did not abuse that yet). So I logged in as the main client user and deactivated the domain with the abusive script, warn the client that they should contact me for information. I want to educate them and make sure they do it right before I reactivate their web domain. Email communication should continue so I don't want their email deactivated.


    Problem with this scheme is that they can log-in to LiveConfig and re-activate the web domains. So I looked at the possibility to deactivate their access to the web domain configuration. This is however impossible. So the only way to inhibit users from reactivating their web domain is by deactivating the whole domain including email.


    The feature I request here is that the server administrator can change the Liveconfig access rights of the main client user, the same way the client user can do this for co-users he created. For the co-users namely appears a Save and Delete button. Would be nice when server administrators can also do this for the main client user. If that would be possible, I would deactivate access to domain administration for the related contract and problem solved.

    In the conclusions of this comparison they refer to it as more suitable for small VPSes and LiveConfig seems to be in that same category. Is php_fpm already supported? Will it one day be supported? Or did somebody use it already with LiveConfig? Could you share your experience?


    (I don't mind when you share it in German)

    Here I posted a feature request in the wrong forum. It would need some extensive changes in LUA, too much to do it myself, beside the fact I was not able to correctly override the postfix lua. So please consider it.

    Aren't you able to change / manage simple files?


    Looks like a rhetorical question. Of course I can't. That's why I use LiveConfig. I'm too lazy to maintain all these complex configuration files with their decentralised development and scattered documentation. My memory is limited and I cannot remember all the different configuration keywords and I am lazy enough that I like the idea that LiveConfig takes care of this.

    In my client's situation there is one contract or subscription with two domains. When I look at the access log I cannot distinguish between the two sites because it has entries of both and there is no domain name in the format.


    The bug is that there should be separate log for each domain, only combining www subdomain with it's domain. The same problem is when you use subdomains. The statistics should be different for each sub-domain. Only www. sub domain can share log-file with its domain.


    It is still in time for this client since the second site is still before release data. But as soon as it is released there will be a problem. Her statistics will be invalid and there is no way to see the statistics for the second domain. This option is not even offered by LC! May the statistics was switched off when the second domain was added? Because that was her initial complain that I try to solve now. Please fix this because there is no way to know what is what in a log, specially when the sites are of similar technology.