Hi Everyone! I’m running into an interesting challenge with my Nextcloud deployment so I’ll start with the details and a quick log copy/paste and ask for any suggestions you might have.
I’ve got a server at home with esxi 6.7. I’m running FreeNAS with an HBA passthrough for storage, as well as a couple of SSD disks for various reasons. I’ve tried all sorts of things to let my FreeNAS shares manage the data volume (NFS, CIFS), but it just felt sluggish so I used a 2 TB Samsung enterprise SSD I had lying around and that is my docker server datastore. Once I get everything set up I plan on backing up to FreeNAS so I still has some level of comfort on redundancy.
Moving on, I have Centos7 with an ansible playbook. It builds out basic stuff, required packages, users, permissions, external volumes for the containers, docker, docker-compose, etc. I can go into more detail if needed. It is pushing a traefik image, a linuxsever.io Nextcloud image, a mariadb image, a redis image, a Onlyoffice image. Everything resides on the single datastore and general tests show over 200 MBps in performance on the virtual disk in both directions. I see nothing else wrong, the nextcloud instance is public and fully functional. It’s snappy and seems to work well. The ONLY problem I have with Nextcloud is WebDAV.
When I take a group of pictures from a Windows machine (I have not tested it in other environments yet), and drag them into a folder in the web version of Nextcloud, it begins the transfer and then hangs and eventually says an error occurred multiple times. Logs show:
|Fatal|webdav|Sabre\DAV\Exception\BadRequest: expected filesize 7043609 got 16384||2019-07-16T18:12:05-0400|
|Error|no app in context|Sabre\DAV\Exception\BadRequest: expected filesize 7043609 got 16384||2019-07-16T18:12:05-0400|
I can provide a logfile if it helps. A few things to note: I can copy a few files and it works. I’m not being specific intentionally because I was unable to identify a precise number or aggregate filesize that causes the failure. I have split DNS at home and the same problem occurs inside my network, so it’s not something regarding public traversal. Everything I can find on these Sabre errors is either unrelated (not docker, using LAMP, etc) or has to do with a single file size error and is repeatable and clear in the failure. If I use the app on my iphone to upload pictures, I will randomly get failures and errors about uploads if I upload a large folder of pictures, but I take no action and it corrects itself and uploads the files eventually. Apps and everything else appear to work as intended and are not slow or problematic, so I’m really struggling to understand what could be wrong.