Resolved Could not get Mozilla tls config: cURL error 77 [cannot update plugin et. al.]

[panelfront]

Server operating system version

CentOS Linux 7.9.2009

Plesk version and microupdate number

Plesk Obsidian v18.0.55_build1800230919.07

Hello , today I start to get this error on SSL/Certificate panel , also I cant make curl requests from my server to external https adresses I got curl 77 error . Is it plesk related issue? , can you help me to fix this issue? Thank you.

 

[Bolivir]

Ive also got the same problem, started today.

 

[Jonathan B]

Same problem here, started today.

 

[Bolivir]

Ive got this problem on a website that was performing a cURL request. There i got error 77 also.
Ive changed the ssh access to /bin/sh and now the cURL work perfectly fine. When changing it back to forbidden, it still keeps working like before.

 

[gennolo]

I have the same issue on several servers that uses PHP to perform cURL calls , they all end up with error 77.

It seems that un automatic update of nss package happened overnight causing corruption on its internal DB, that is being used by PHP cURL by default.

A verbose logging of a PHP cURL call produces :

* Initializing NSS with certpath: sql:/etc/pki/nssdb
* Unable to initialize NSS database
* Initializing NSS with certpath: none
* Unable to initialize NSS
* Closing connection 2 HTTP Code: 0

 

[Jonathan B]

Thank You Bolivir,
Now it works fine, changing ssh access to /bin/sh.

 

[panelfront]

Ive got this problem on a website that was performing a cURL request. There i got error 77 also.
Ive changed the ssh access to /bin/sh and now the cURL work perfectly fine. When changing it back to forbidden, it still keeps working like before.

I think this is not a global fix for this issue , did anybody managed to fix this issue?

 

[Peter Debik]

After the operating system's update of NSS files, a restart of PHP-FPM services is required to mitigate the issue. It is not necessary to change the shell setting. However, by changing the shell setting, likely that restart also occured, for that reason it works after setting it to "forbidden" again.

--> Solution is to restart all PHP-FPM services.

 

[Bolivir]

Thanks Peter! Will try that, @panelfront my comment was not meant to be "a permant fix", rather a it works now fix but dont know why .

 

[kevinjansen]

How do you restart if you use the dedicated php-fpm? Since the primary php-fpm 8.2.11 is not setup at tool-settings > services

 

[kevinjansen]

Doing a:
service sw-engine restart && service sw-cp-server restart
is enough for the LetsEncrypt part in Plesk panel.

 

[Peter Debik]

According to further analysis by staff we also think that a restart of sw-engine could be necessary.
New KB article: https://support.plesk.com/hc/en-us/articles/18463826337815

 

[Cut the Mustard]

We had a similar issue and upgraded to 18.0.56 and restarted the servers and this seemed to solve the issue.

 

[GabyR]

Updated Plesk to 18.0.56 and restarted the server and all the issues are resolved. Thanks!

 

[D4NY]

According to further analysis by staff we also think that a restart of sw-engine could be necessary.
New KB article: https://support.plesk.com/hc/en-us/articles/18463826337815

Same here, solved with this!

 

[Ludo]

Same here, solved with this!

Solved with this too!

 

[Rot]

Problem solved in my servers, thanks!

 

[Deagan]

Anyone experiencing ipv6 issues? Running CentOS 7.9 with Virtuozzo. From the host I can ping6 (internet+local containers). From the containers/vm's I can ping the host but I can't ping external. The issue started after the yum update in all vm's ( also had the curl/no resolve possible from php curl issue ).

No settings have been changed, I've already checked all settings I could find ( forwarding, accept_ra, added/removed ipv6 ip from container, added new route to ipv6 gw in container, restarted network in container etc. ).

Any help would be really appreciated.

 

[Peter Debik]

@Deagan This is unrelated to the topic of this thread. Please open your own thread for this new question.

 

[knattk]

According to further analysis by staff we also think that a restart of sw-engine could be necessary.
New KB article: https://support.plesk.com/hc/en-us/articles/18463826337815

Solved! Thank you.