DirectAdmin v1.645 has been released

[japper83]

@ZipperZapper

We have the same problem. After the updates last night the admin back-ups fail on two servers. With this error:

ftp_upload.php exit code: 67
ftp_upload.php output: curl: (67) Access denied: 504
curl return code: 67

We are using FTPS for the back-ups.

We are running:
Directadmin: 4c6940b34d995a6418287b8b89b4471eb07a9c19
OS: CentOS Linux 7

 

[dafang]

The system has been working flawless for years. It started breaking recently with the v1.645 wave of updates.
The most recent version did not fix it. On the other end of the line, nothing has changed.

What steps can I take to fix this?

Does the error go away if you revert to "stable" release 1.644 ?

 

[ZipperZapper]

If I install 1.644 and run the backup it still breaks, but with a different error:

ftp_upload.php exit code: 35
ftp_upload.php output: curl: (35) error:1409442E:SSL routines:ssl3_read_bytes:tlsv1 alert protocol version
curl return code: 35

 

[fln]

@japper83 access denied error is not related to the DA update, please check your credentials.

@ZipperZapper I could dig deeper into the issue if you would give me SSH access to your server via private message, our SSH key is https://directadmin.com/authorized_keys

 

[japper83]

We fixed our problem. Turns out CURL was updated with the OS. We reverted to an older version of CURL. Now it's working.

 

[Active8]

We fixed our problem. Turns out CURL was updated with the OS. We reverted to an older version of CURL. Now it's working.

Hi can you elaborate more ?
Wich OS ? which version cURL ?
DA decided to stop with cURL updates trough Custombuild since then its an mess.
As far i can see I don't see any cURL updates in Almalinux 8.x (ver 7.61.1 in use)

 

[zEitEr]

bash -x /usr/local/directadmin/custombuild/build doMigrateToSystemCurl

By the way the function doMigrateToSystemCurl might need to update users crontasks. Something like the following:

perl -pi -e 's#/usr/local/bin/curl#/bin/curl#' /var/spool/cron/*

Otherwise crontasks might stop working.

 

[ZipperZapper]

@zEitEr You're correct. That's how I first noted something serious had changed, one of my websites heavily depending on curl crontasks came crashing down. It really is a very serious change.

Although in my case the curl actions are inside bash files called by cron, so hard to cover everything for users.

 

[zEitEr]

proftpd[1751162] xx.xx.xx.xx : mod_sftp/1.1.1: Server 'ProFTPd': cannot support both FTPS (TLSEngine on) and SFTP (SFTPEngine on) for the same host.

For those who already installed proftpd.conffrom GitHub using the guide from https://help.poralix.com/articles/proftpd-cannot-support-both-ftps-and-sftp-for-the-same-host please re-install the config file. An error was found in it, and it is fixed now.

 

[Daniel_Dog]

When will this update come to stable?

 

[fln]

@realcryptonight, when we release DA 1.646 into current, stable will be bumped to 1.645.

We expected to release 1.646 before the holiday season, but we will see how it goes.

 

[japper83]

Hi can you elaborate more ?
Wich OS ? which version cURL ?
DA decided to stop with cURL updates trough Custombuild since then its an mess.
As far i can see I don't see any cURL updates in Almalinux 8.x (ver 7.61.1 in use)

On the servers with the back-up problem we are running Centos 7.
The server updates curl to version 7.86. Then the problems started. We downgraded curl with a hand compiled version of 7.61.

 

[Richard G]

On the servers with the back-up problem we are running Centos 7.

Custombuild curl would be replaced by the original one of the OS by DA. Seems that did not happen.

For Centos 7 the OS version is currently 7.29.0 which you should be able to install via yum.

 

[brandyou]

Hello,
something wrong with Apache + PHP fastcgi

At DirectAdmin v1.645 after Apache update or recompile (./build apache n), PHP scripts don`t work, domain error log shows:


/var/log/httpd/suexec_log shows:



#suexec -V is missing "-D AP_SAFE_DIRECTORY="/usr/local/safe-bin", current output:



If we change DirectAdmin version back to v1.644 (stable) and recompile Apache, everything works fine. So I think v1.645 have mistake at suexec patch.

We are facing this same issue at DirectAdmin version 1.645 and some site still down. Any solutions other than reverse back to v1.644 from v1.645?

 

[fln]

We are releasing new DA 1.645 build c52eab63c147d7c6ca97405d0b87272465465d20, it addresses the following issues:

• CB fix for Apache + PHP suexec, thanks @brandyou
• Backup upload via FTPS fix for Ubuntu 18, thanks @ZipperZapper
• Couple of regressions in Evolution File Manager - showing English instead of custom language, tree directory expansion refresh loop

 

[IXPLANET]

@realcryptonight, when we release DA 1.646 into current, stable will be bumped to 1.645.

We expected to release 1.646 before the holiday season, but we will see how it goes.

Hai fln , in 1.646 version when i was logging in chrome on my phone , a left menu or panel menu is missing / dissapear , but if switch to dekstop mode in chrome mobile, left menu show .

 

[Erulezz]

@fln I still see   in the Server Statistics page as reported in this post: #86

 

[Ohm J]

@Erulezz
that's " ", come from 1.643

DirectAdmin v1.643 has been released

autocomplete='off' spellcheck='false' autocorrect='off' autocapitalize='off' @fln should have option to set these on any textbox these 4 options will help to solve problem on Iphone device. that somecase it auto modify text when input. im web developer, so Iphone have problem on input field if...

forum.directadmin.com

but it still not fixed.

 

[Zhenyapan]

export DA_INTERACTIVE_CUSTOMBUILD=yes
didn't work anymore?
someone told me that it works only with:
export DA_CHANNEL=stable
but if I want "current" - I must install stable, then change to current and rebuild all again?

 

[fln]

An update (build e20a65fa9b8c6a70c07a3b0a8962320dbb4bde6b) is released to fix for login page that prevents login when security questions are enabled but no questions are added.