Resolved Website subscription cannot use cURL results in cURL error 28: Resolving timed out

[Kaspar]

Sorry you're still having issues.

1. What setting do I need to allow outgoing cURL requests or specify what domains are allowed? I can only whitelist ports and ip's.

Are unable to reach any url when using cURL? Of do some work and others not?

 

[TeunieP]

When I try to reach urls using cURL from SSH I don't have any issues but Wordpress sites keep reporting connectivity issues about connecting to external servers.

Currently I have spent so much time on this issue I am thinking about reinstalling the entire server.
All these performance and connectivity issues came out of nowhere so I still believe an update must have caused this. But I don't seem to find any usefull information in the logs...the server is just very slow in running Wordpress sites although there are enough unused resources. It's almost like it is throttling down subscription performance for no reason...

 

[TeunieP]

I also have a Joomla website that uses a Gantry template which is suddenly very slow as well in loading pages. And if I switch to a different template it is solved..but there is a cause behind all this. The server used to be one of our fastest before 1 month ago.

 

[TeunieP]

P.s. the server CPU load is about 0.2 on a 8-core cpu, 37% physical ram and 73% physical disk usage. FYI.

 

[TeunieP]

1 migrated one of the problematic sites to a different server also running the same version of Centos and Obsidian and there are no issues here...

 

[TeunieP]

Issue came back, now also on the second server. Painfully slow page loads, Wordpress backend is even worse. Installing plugins is nearly impossible. Curiously though not all Wordpress sites have the same issues.

Example (waiting for almost 6 seconds);

 

[TeunieP]

Also look at this, 2 different wordpress sites. One is painfully slow other is quick. However the designer who build the slow site says it is running fine in a DirectAdmin server. So must be Plesk related right?

curl -s -w '\nLookup time:\t\t%{time_namelookup}\nConnect time:\t\t%{time_connect}\nSSL handshake time:\t%{time_appconnect}\nPre-Transfer time:\t%{time_pretransfer}\nRedirect time:\t\t%{time_redirect}\nStart transfer time:\t%{time_starttransfer}\n\nTotal time:\t\t%{time_total}\n' -o /dev/null https://site1.com
Lookup time:            5,514
Connect time:           5,516
SSL handshake time:     5,634
Pre-Transfer time:      5,634
Redirect time:          0,000
Start transfer time:    5,635

Total time:             5,635

and other wordpress site;

[root@websrv ~]# curl -s -w '\nLookup time:\t\t%{time_namelookup}\nConnect time:\t\t%{time_connect}\nSSL handshake time:\t%{time_appconnect}\nPre-Transfer time:\t%{time_pretransfer}\nRedirect time:\t\t%{time_redirect}\nStart transfer time:\t%{time_starttransfer}\n\nTotal time:\t\t%{time_total}\n' -o /dev/null https://site2.com
Lookup time:            0,028
Connect time:           0,029
SSL handshake time:     0,144
Pre-Transfer time:      0,144
Redirect time:          0,000
Start transfer time:    0,178

Total time:             0,180

 

[TeunieP]

I also tried these steps but does not help.

https://support.plesk.com/hc/en-us/articles/213940705

 

[Bitpalast]

Maybe you can try to change the nameservers that your server uses to Google servers, e.g.
nameserver 8.8.8.8
nameserver 8.8.4.4
in /etc/resolv.conf

 

[TeunieP]

Hi,
I checked and I am already using 1.1.1.1 now changed to your suggestion and a reboot, but this has not changed anything.
Still around 5 seconds of waiting time before the pages are loaded.

 

[Bitpalast]

1.1.1.1? That does not read like 8.8.8.8 or 8.8.4.4. According to your measurements you are suffering a name resolution problem. It could really help to switch nameservers. At least you could rule this possibility out by doing so.

 

[TeunieP]

Hi Peter, 1.1.1.1 is Cloudflare open DNS which should be fine. However I think you didn't read my message properly because I wrote "now changed to your suggestion and a reboot, but this has not changed anything" meaning I changed the DNS to your suggestion 8.8.8.8 and 8.8.4.4 and performed a reboot after but this did not help, unfortunately.

 

[TeunieP]

Also, the first line in the resolv.conf is;

search defaultdomain.com

Where defaultdomain.com is the default domain I set-up in Plesk so I assume this is right? This was never changed in the last 5 years.

and then
nameserver 8.8.8.8

the nameserver 8.8.4.4 seems to be removed after the reboot automatically

I can also use ping command to lookup external sites like google.nl without issues.

 

[TeunieP]

Hi,
I had to use;
chattr +i /etc/resolv.conf

to forcefully keep this file from being edited.
After removing search defaultdomain.com it appears to have solved the issue with the 5sec delay.

Curiously all our Plesk servers were affected by this issue.

 

[TeunieP]

At first it looked like there was an improvement, but the delay is still the same with a lookup time of 5,5 seconds...

 

[TeunieP]

This was the solution thanks to Plesk support through Open Provider;

As a workaround, to enforce single-request queries, I added the following line in the /etc/resolv.conffile:

options single-request

 

[Kaspar]

Good to hear you've got a (workaround) solution. And thank you sharing the solution here. This might benefit other users as well. I myself, for example, did not know about the options single-request configuration. Which seems very useful.