I’m on NextCloud 10.0.1, lighttpd, mysql, debian.
dbfiles, www-home and data dir is mounted on ext4 (external disk)
The overall performance of Nextcloud in the webinterface is pretty fine (about 2-3 secs for a page, sometimes 10) The same goes for webdav access - it’s a tad slower but still in the seconds.
I have one folder with about 5000 pictures - and only pictures. All around ~5MB. No subfolders.
Opening that folder in webdav takes about 20-30sec - which is Ok and something which is within my expectations.
But opening the same folder via the web-interface takes about 4+ minutes. (https://example.com/index.php/apps/files/?dir=/Photos/Camera roll)
When doing this request I see, that iotop is idling and mysql process is idling. Top shows me, the lighttpd process is running fast-cgi at about 30-40% CPU usage the full 4 minutes.
When doing a SHOW PROCESSLIST in mysql I see nothing but a few short bursts with the SQL select commands which are usually too fast to note. The slow-query.log is empty.
I was switching from Apache2 to lighttpd recently because I had the same issues on Apache2 and was under the impression, things would become better when using lighty. I was wrong…
Does anyone has an Idea, where I could look further? It seems, that the code used by the webinterface is doing some extra stuff, the webdav code doesn’t.
And really, this is the first time in my entire life, that webdav is faster then a plain http request…