XF 2.2 Importer stops with 503 error

[Black Tiger]

Service Unavailable​

The server is temporarily unable to service your request due to maintenance downtime or capacity problems. Please try again later.

I was importing a test version of IPB 4.4.10 to Xenforo 2.2. It started fine and then suddenly I got the above.

Both were protected by .htaccess but my ip and 127.0.0.1 were both allowed.
Server cpu Intel(R) Core(TM) i7-8700 CPU @ 3.20GHz
At this moment 64 GB Ram and at least 35 GB ram free.

It happened during importing threads.
I reloaded the page and at this moment it's going further, but I would really like to know what is causing this and how to prevent this on the real import.

 

[Mike]

An error like this is coming from your host/web server configuration. You'd need to get advice from them on whether it's preventable. (Presumably it's automatically triggered based on the number of requests or load/CPU time used, or I suppose it could be an anti-DoS configuration.)

Using the CLI importer would probably avoid it.

 

[Black Tiger]

An error like this is coming from your host/web server configuration.

I am the hoster. And we don't have any limits on load/cpu time.

It's all being imported on localhost so it shouldn't have been anti-DOS. Maybe php-fpm went out but I didn't found anything about that in the php-fpm logs.

CLI importer? Where can I find that and a manual to it? Do I need this if all data is local on the server?

 

[Paul B]

https://xenforo.com/docs/xf2/importing/#cli-importer

 

[Black Tiger]

Thanks, if it happens again I will use the cli importer.

 

[Black Tiger]

Odd. It just happened again, I linked to my testforum for my moderator and got the Service Unavailable again.
When trying to visit myself, only white screens were presented.

That took a little time and then suddenly everything worked well again.

In the domain error log I've seen this:

[Wed Jan 06 15:50:17.080472 2021] [proxy_fcgi:error] [pid 3740:tid 140136259680000] (32)Broken pipe: [client 84.xx.xx.xx:51225] AH01075: Error dispatching request to : (sending stdin),

This striks me as odd, because it looks like it happened because I linked from my live IPB forum in a private message, to my Xenforo test forum.
The test forum is restricted with .htaccess but 127.0.0.1 is allowed, also the ip of me and my moderator.

Never had this before with anything so I really don't know where to look.
Nothing odd in the main apache logs or the php-fpm73 log.

 

[Mike]

Is that the only log entry that might be related? I'm really not sure what may be happening here. I have found some references to this error, though I'm not sure if they're relevant (particularly when it's a sporadic error). It seems like a fairly low level error between mod_proxy_fcgi and php-fpm -- I doubt that PHP code is actually executing at the time the error happens. I've seen some references that relate to request sizes that are too large, but that shouldn't be relevant unless this happened when uploading a file.

 

[Black Tiger]

This is the one in the domain error log. The apache error log did not show anything.
There was a reference, I can post the full line of the error log, but it was the only thing in the error.log of the domain:

[Wed Jan 06 15:50:17.080472 2021] [proxy_fcgi:error] [pid 3740:tid 140136259680000] (32)Broken pipe: [client 84.xx.xx.xx.:51225] AH01075: Error dispatching request to : (sending stdin), referer: https://www.mydomain.org/profile/26-mymoderator1?wr=eyJhcHAiOiJm... etc... k5OH0=

That is the full line. It's from when I send the pm from my original forum to my mod and here the check of the existing link was done as far as I can see.

I've seen a 503 at the same time in the access log of the domain, but that is because I was writing the pm at that moment to my moderator, when posting the link there is a check done automatically and this could not find the url.
So it wasn't when uploading a file.
After that I had a white screen on my xen test installation, even when refreshing a couple of times, also when trying with Chrome.
Very odd.

This is from the domain's access.log when I got the white screen when testing with Chrome and tried with refresh:

84.xx.xx.xx - - [06/Jan/2021:15:54:33 +0100] "GET /xentst HTTP/2.0" 301 394 "-" "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/87.0.4280.88 Safari/537.36"
84.xx.xx.xx - - [06/Jan/2021:15:54:33 +0100] "GET /xentst/ HTTP/2.0" 200 99 "-" "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/87.0.4280.88 Safari/537.36"
84.xx.xx.xx - - [06/Jan/2021:15:54:35 +0100] "GET /xentst/service_worker.js HTTP/2.0" 304 58 "https://www.mydomain.org/xentst/service_worker.js" "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/87.0.4280.88 Safari/537.36"

I've seen that last line more often in the log this way.