krajenski

Tacho! Lag wohl hier dran: Im Non-SSL-Zweig drin gewesen, aber nicht im SSL-Zweig: Das Markierte ist nun auch drin. Ich schätze mal dass das früher ggf. einmal nur ein zentrales Setting war? Nur eine Vermutung.

Sure, will do!

Allright, I think I can sum this issue up to this misbehaviour of the DNS editor after all: Setting a CNAME record for "www" produces a duplicate entry in the zone, which leads to it not being loaded. (The other "@" issues seems like some non-RFC-conformance thing).

Update #2: And it does not work really. The zone is generated, but not loaded. Oct 23 13:44:26 server named[8798]: dns_master_load: /etc/bind/domains/domain.de.zone:8: domain.de: CNAME and other data Oct 23 13:44:26 server named[8798]: zone domain.de/IN: loading from master file /etc/bind/domains/domain.de.zone failed: CNAME and other data Maybe it is not possible to set a CNAME record for the TLD ("@") itself. Never needed this before. Will look further into it.

Update: Now I was able to set the "@" CNAME-Record for my "case A", but not in a way that makes sense... Being desperate, I experimented with deleting the "@ MX" Record (!), and after that, there was no "conflict" anymore and the end result looks like this. I fortunately was able to do this in this productive case, because, luckily, the MX works also with the implicit default entry in this case, phew (Some customers have MX-Records not on our server, where this move would not have been possible) So some bug, after all 🙂

Then try my suggestion in line No.3: https://www.veeam.com/linux-backup-free.html Key Bulletpoints: - Bare-metal restore - Back up an entire Linux system or specific files (crash-consistent) - Console UI or command line - Multiple Jobs - Multiple Backends (local, NAS, Cloud) - Encryption: So it doesn't matter if backing up to 3rd-Party - and so so on... Try it out!

Thanks for your quick reply. To answer in full: It is configured... ...but not in productive use (yet). -> Changed thecode -> Tested + Works for me! No Warnings.

Hi! Yes, name resolution in general is fine on this machine. After looking at some articles like this: https://bugs.php.net/bug.php?id=73149 https://stackoverflow.com/questions/36280957/dns-get-record-a-temporary-server-error-occurred https://dustri.org/b/a-short-tale-on-phps-dns_get_record.html ...I included a 'var_dump(' ... in the lines that were mentioned in the error message(s). (226 + 228 + 229) of .../froxlor/lib/Froxlor/PhpHelper.php and did a: # /usr/bin/php -q /var/www/html/froxlor/scripts/froxlor_master_cronjob.php --run-task 4 --force PHP Warning: array_merge(): Expected parameter 1 to be an array, null given in /var/www/html/froxlor/lib/Froxlor/PhpHelper.php on line 229 NULL PHP Warning: Invalid argument supplied for foreach() in /var/www/html/froxlor/lib/Froxlor/PhpHelper.php on line 234 [warning] Skipping Let's Encrypt generation for some-customer-domain.de due to no system known IP address via DNS check array(0) { } array(1) { [0]=> array(5) { ["host"]=> string(22) "www.some-customer-domain.de" ["class"]=> string(2) "IN" ["ttl"]=> int(10799) ["type"]=> string(1) "A" ["ip"]=> string(14) "109.237.140.10" } } PHP Warning: array_merge(): Expected parameter 1 to be an array, null given in /var/www/html/froxlor/lib/Froxlor/PhpHelper.php on line 229 NULL PHP Warning: Invalid argument supplied for foreach() in /var/www/html/froxlor/lib/Froxlor/PhpHelper.php on line 234 [warning] Skipping Let's Encrypt generation for www.some-domain.de due to no system known IP address via DNS check array(0) { } array(1) { [0]=> array(5) { ["host"]=> string(13) "some-domain.de" ["class"]=> string(2) "IN" ["ttl"]=> int(10493) ["type"]=> string(1) "A" ["ip"]=> string(14) "109.237.140.10" } } PHP Warning: array_merge(): Expected parameter 1 to be an array, null given in /var/www/html/froxlor/lib/Froxlor/PhpHelper.php on line 229 NULL PHP Warning: Invalid argument supplied for foreach() in /var/www/html/froxlor/lib/Froxlor/PhpHelper.php on line 234 [warning] Skipping Let's Encrypt generation for some-domain.de due to no system known IP address via DNS check array(0) { } (...and so on...) I'm not entirely sure how to intepret it. PS: If I run this multiple times and 'diff' the output file, the result is identical. It fails for like ~25 entries, of the many hundreds in our system. Anyway, I can't see a production or productive impact of this at the moment. Greets, Sebastian

Hello there! Just my 2 cents of experience here. We follow a two-sided approach here. For having a consistent bootable desaster-backup for "*.*" we use this: https://www.veeam.com/linux-backup-free.html The base product is free. To quote d00p: There are so many backup solutions out there to choose from. On the other hand, for having some consistent database-backup rotation on the machine itself at hand for quick access and restore, I scripted this: (Run by crontab which has 'MAILTO=' defined (important in my eyes in general ! ) #!/bin/bash ############################# export login=debian-sys-maint export password=`grep password /etc/mysql/debian.cnf |tail -1|cut -d"=" -f2|sed 's/ //g' ` ############################# PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin find /localbackup/ -type f -mtime +3 -iname "*gz" -exec rm {} \; mkdir -p /localbackup day=`date +%Y%m%d` echo $day - Starting Backup of all current local databases mysql -s -r -u $login --password="$password" -e 'show databases' -N | while read dbname; do echo Work at - $dbname mysqldump -u $login --password="$password" --complete-insert --single-transaction "$dbname" | gzip > /localbackup/$day-"$dbname".sql.gz; done echo Done -> Depending on the actual distribution you use this may need some adjustments of course. For me it was important to have something with no included fixed credentials -> Result: All DBs in existing, compressed in one directory, with a (in this case) +3 day rotation. Greets, Sebastian

I understand, sure, there you go: ------------------------------------------------------- "Edit domain"-settings for Case "B" (the good case) are: - No alias domain - No subdomain of a full domain - Allow editing: yes - Next two fields: empty - Documentroot: /var/customers/webs/msolutions/eicar.eu/ - IP: <IPV4>:80 - ServerAlias value: WWW - SepLog: No - Own V-Settings: Empty - Apply special: yes - Write an...: yes - Write an...: yes - SSL: yes - SSL-IP: <IPv4>:443 - SSL redirect: off - LetsEncrypt: yes - Override...: no - ProtVersions: TLS1.2 only - Ciphers: our default string... ECDHE ... and so on - Next two: empty - Include non-ssl: Off - HSTS: 0 - Include HSTS: off - Include domain HSTS...: off - OSCP: off - Honor...: Off - OpenBase: yes - PHP enabled: yes - PHP config: 7.3 - apply php-config to all: yes - Next two: empty - Nameserver: Yes - Zonefile: empty field - Emaildom: yes - onlymail: off - sub as email: Never -> Additional info: This TLD is quite old an has been in the system for some year(s). -------------------------------------------------------------------------------------- "Edit domain"-settings for Case "A" (the bad case) are: - No alias domain - No subdomain of a full domain - Allow editing: yes - Next two fields: empty - Documentroot: /var/customers/webs/the/customer/dir - IP: <IPV4>:80 - ServerAlias value: WWW - SepLog: No - Own V-Settings: Empty - Apply special: yes - Write an...: yes - Write an...: yes - SSL: NO (all subsequent SSL-fields empty or implicit default, no checkboxes checked) - OpenBase: yes - PHP enabled: yes - PHP config: 7.3 - apply php-config to all: yes - Next two: empty - Nameserver: Yes - Zonefile: empty field - Emaildom: yes - onlymail: off - sub as email: Never -> Additional info: This entry is only a few days old and rather newly in the system.

Hi there! As the following issue is not affecting all our domains in the DNS editor, I post it here before filing a bug in github. We have an two-parted issue with adding CNAME-records to DNS zones using the Froxlor DNS-Editor. I have for example two DNS zones here, which behave very differently when adding CNAMES. -> Please see attached two pictures as information basis Case A: Problem 1) When adding an "@" CNAME to this zone this happens: There already exists a resource-record with the same record-name. It can not be used as CNAME. Problem 2) When adding an "www" CNAME entry to this zone this happens: The entry gets duplicated -> TWO "www" entries exist, and the zone is not loaded due to error. The "custom" enty is not respected in the usual manner: It should be replaced. Case B: - Both things simply work like a charm Neither do I see what I am doing wrong, or in which way there are different that this may happen. I even had a look into 'domain_dns_entry' table, which did not reveal a conclusion for me here. Froxlor Version: 0.10.21 Many thanks in advance, Sebastian

Hi there, I'm looking for a hint on how-to debug the following output from "/usr/bin/php -q /var/www/html/froxlor/scripts/froxlor_master_cronjob.php --tasks": PHP Warning: dns_get_record(): A temporary server error occurred. in /var/www/html/froxlor/lib/Froxlor/PhpHelper.php on line 226 PHP Warning: dns_get_record(): A temporary server error occurred. in /var/www/html/froxlor/lib/Froxlor/PhpHelper.php on line 228 PHP Warning: array_merge(): Expected parameter 1 to be an array, bool given in /var/www/html/froxlor/lib/Froxlor/PhpHelper.php on line 229 PHP Warning: Invalid argument supplied for foreach() in /var/www/html/froxlor/lib/Froxlor/PhpHelper.php on line 234 PHP Warning: dns_get_record(): A temporary server error occurred. in /var/www/html/froxlor/lib/Froxlor/PhpHelper.php on line 226 PHP Warning: dns_get_record(): A temporary server error occurred. in /var/www/html/froxlor/lib/Froxlor/PhpHelper.php on line 228 PHP Warning: array_merge(): Expected parameter 1 to be an array, bool given in /var/www/html/froxlor/lib/Froxlor/PhpHelper.php on line 229 PHP Warning: Invalid argument supplied for foreach() in /var/www/html/froxlor/lib/Froxlor/PhpHelper.php on line 234 It probably has something to do in context with: "gethostbynamel6" function in there. We're not actively/productively using IPV6 (yet). The only thing we have (so far) configured it is in the menu item "IPs and Ports" for some tests in the past. But there should be no BIND-zones with IPV6. The occurence of this issue is new. It came up after the major update from 0.9x to 0.10.x Best regards, Sebastian

...and all went well this way! Thanks for your superb support.

Great, thanks. I hoped for an alternative method of updating in existing Could you give me a quick hint on how the update is done in the shell? Or a quick link to the documentation in this matter.

Good Morning d00p, many thanks for your fast reply and the work you do! I just tried it again, the answer is: 201912313 PS: I now dumped the failed db for possible further analysis before restoring. Best Regards, Sebastian