• If you are still using CentOS 7.9, it's time to convert to Alma 8 with the free centos2alma tool by Plesk or Plesk Migrator. Please let us know your experiences or concerns in this thread:
    CentOS2Alma discussion

Question I want to downgrade from php7.3.20 to 7.3.19

Z

Zoo3

Regular Pleskian
I updated to PHP 7.3.20 and faced an issue with Nextcloud not working and I did.
The user can log in to Nextcloud, but immediately after that a 502 error occurs.

[error] 21578#0: *2 recv() failed (104: Connection reset by peer) while reading response header from upstream, client: IP-ADDRESS, server: MY-NEXTCLOUD, request: "POST /index.php/login HTTP/2.0", upstream: "fastcgi://unix:///var/www/vhosts/MY-NEXTCLOUD/php-fpm.sock:", host: "MY-NEXTCLOUD", referrer: "https://MY-NEXTCLOUD/"
Click to expand...

I once uninstalled PHP7.3.20, and then reinstalled PHP7.3.20 I'm trying.

I want to downgrade to PHP 7.3.19. Please tell me how to do that.
 
I've switched to PHP 7.4 and am able to use Nextcloud. I'm not sure if it's a problem with NC or PLESK-PHP7.3.
 
Hi,

I have the same issue after the upgrade of PHP 7.3.
I've done what @Zoo3 said, install PHP 7.4 and run the NC on there, works like a charm! Thanks!

Greetings,

Bart
 
Hi,

just in case a Plesk employer sees this thread: Not all of us get offered PHP 7.4 (Onyx Users, for example, that can't upgrade to Obsidian) and therefor a fix for PHP 7.3 is necessary.
Over in the nextcloud forum there are several people having this issue, however, only with Plesk PHP 7.3.20, not with regular PHP 7.3.20.
 
No,i just log in and get this 502 error ... i try on different customers, and nextcloud versions. same everywhere
 
Has anyone in this case tried to
# service httpd restart
(or on Ubuntu, Debian etc. # service apache2 restart)
# service plesk-php73-fpm restart
?
Occasionally, FPM service "disconnects" from Apache when it thinks, other children or instances need to finish a task before a new one can respond to Apache. The issue might also have something to do with the pm.max_children and pm.max_requests settings. Try pm.max_children=25, pm.max_requests=10000 for testing purpose to avoid frequent hangs of the PHP-FPM service.
 
It seems that neither from this, nor the other thread anyone has filed a bug report. I cannot file that report, because I cannot reproduce the issue. It might be a good idea if somone did that so that it can be forwarded to developers:

Reports

Please, click Post New Thread button and fill the form if you really believe you discovered a new bug and need to report it to the Plesk developers. It does allow us to handle and resolve your requests much more efficiently. Please be careful and do not submit any found problem with this...
talk.plesk.com talk.plesk.com
 
I'd not use "dynamic" but "on demand", which leaves the calculation how many instances are needed to the system. The "max_requests" should be much higher, because else, children need to die and re-created frequently. Not sure about the additional opcache-settings either. It seems that none of our customers uses these.

However, I cannot verify the opcache issue described in the other thread or here. Still suggesting to file a report with Plesk.
 
i change to "on demand", don't work too :-(
But we buy a new server with centos 8, i start migrate on thursday
 
Please read the limitations on Plesk on CentOS 8 here:
support.plesk.com

CentOS 8/RHEL 8 support in Plesk

Applicable to: Plesk for Linux Question Does Plesk support CentOS 8/RHEL 8? Answer Support for CentOS 8/RHEL 8 is available starting from Plesk Obsidian 18.0.25. Known limitations: No live/i...
support.plesk.com support.plesk.com
 
I cannot give advice on that, just pointing out that there are limitations for CentOS 8 with Plesk.
 
  • Like
Reactions: BNO
Peter Debik said:
I'd not use "dynamic" but "on demand", which leaves the calculation how many instances are needed to the system. The "max_requests" should be much higher, because else, children need to die and re-created frequently. Not sure about the additional opcache-settings either. It seems that none of our customers uses these.

However, I cannot verify the opcache issue described in the other thread or here. Still suggesting to file a report with Plesk.
Click to expand...
This is my report with Plesk, isn’t it?

Edit: no, it isn’t. sorry, am on mobile atm. My Thread is my report with Plesk, isn‘t it?
 
Here is the link again, where bugs can be reported:

Reports

Please, click Post New Thread button and fill the form if you really believe you discovered a new bug and need to report it to the Plesk developers. It does allow us to handle and resolve your requests much more efficiently. Please be careful and do not submit any found problem with this...
talk.plesk.com talk.plesk.com
 
thanks, sorry, was on mobile and I didn't see it. I will reference this thread, my thread and nextcloud forum threads.
 
You must log in or register to reply here.

Similar threads

M
Issue End of script output before headers / (104)Connection reset by peer: [client xx.xx.xx.xx:48] mod_fcgid: error reading data from FastCGI server
Replies
17
Views
1K
Peter Debik
P
B
Issue Error upstream sent too big header
Replies
7
Views
2K
BNO
B
Kulturmensch
Resolved 502 Bad Gateway after upgrade to Obsidian 18.0.40
Replies
17
Views
3K
Kulturmensch
Kulturmensch
Jorge Batres
Issue Symlink between to domains in same server not working
Replies
6
Views
1K
jmar83
J
M
Question Errors during loading homepages
Replies
1
Views
1K
Bitpalast
Bitpalast
Back
Top