New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We鈥檒l occasionally send you account related emails.
Already on GitHub? Sign in to your account
MOVING files already present in NC to new, encrypted folder causes problems #2519
Comments
@misch7 @camilasan This issue seems to be isolated to the Desktop client. Can you reproduce it? |
Is the desktop client version 3.0? |
Yepp. |
Could you provide desktop client logs? Please run the client with --logdebug |
Also If you ever tried end-to-end encryption before version 3.0, please follow the instructions from this post: to clean up the keys as that can break the functioning of >3.0 if you had a malformed key. |
As soon as E2E app is enabled desktop clients (3.0.0 / 3.0.1) sync like forever. Before enabling e2ee sync took few seconds, now one iteration doesn麓t matter if and how much to sync takes at least 5 minutes. Significant load on the server (PHP-FPM and MYSQL). Even when the desktop client is closed server is still busy. Looks like there麓s an issue in terms of communication and server tries to satisfy huge mysql select requests (I guess on the filecache table). iOS client sync works instantly, nevertheless the nextcloud/desktop guys pointed us over here (issue seems to be part of the server component). Basically e2ee isn麓t usable currently so the "big bang blog announcement" ("production ready e2ee!") is far away from being true. |
I experienced similar issues:
Unfortunately I did not find time to really troubleshoot in depth and poke through logging of all components, that's on my to do list. I disabled and uninstalled the E2E app in the server, but this load still remained, also after reboot of all components. But one thing I noticed during this increased load, there also was a substantial load increase (cpu and traffic) on my samba server. edit: in the mean time I upgraded from NC 19.0.1 to 19.0.2, above behavior in this comment is not reproducible anymore. |
I've observed the same, huge load on php-fpm and database (postgres). The following SQL statements are executed in the database, within miliseconds, millions of those(!):
I hope it helps to debug the issue. As soon as I disable E2E on the server, things are stable as expected. |
I see similiar SELECT commands occupying the MySQL (and PHP-FPM) for a loooooooong time. What next? |
These are the access.log lines when a nextcloud client starts up. Not too much, I think. The postgresql.log at the same time, |
No activity... I can confirm this issue still persists with latest desktop client v3.0.2 and current Nextcloud v19.0.3.1. |
This is desktop related, so I move it to desktop repo. |
@tobiasKaminsky got a couple of duplicates of that one already, actually I think it was good this one was on the e2ee app repository. Yes the desktop client is apparently revealing something but to me there are strong chances the bug is on the server side (could be wrong of course, but so far with all the data points I got I couldn't pin point it to something concrete on the client end). What makes me think this is that very often in those cases, the client isn't doing much apart from waiting for a PROPFIND (in one of the reports I got it basically times out after a while), that can also be seen in some of the comments here where someone mention nothing happens in the client debug log but there's a storm of SQL requests going on server side. Clearly the client issued a command and waits for an answer that never comes because something went haywire on the server. I'd like someone with knowledge on how that components work to actually look into it. If I'm right that'd explain all the "sync gets stuck when e2ee is in the mix" and "server load massively increased" reports which I got so far. |
Where to find someone with that specific knowledge - isn't this the place to look for one? Unfortunately I'm struggling with it too from a users perspective while at the same time not being able to have a "developers view" on it. |
How to use GitHub
Steps to reproduce
Expected behaviour
Data encrypted and sync'd to server
Actual behaviour
No sync to server.
Suddenly system got unresponsive
Found that many MySQL processes run at full whack --> needed to restart mysql server --> server responsive again
Tried to delete the encrypted folder --> got a 403 message displayed in client --> disabled e2ee --> deleted folder
Server configuration
Operating system:
Ubuntu 20.04.1 aarch64
Web server:
Nginx 1.18
Database:
Mysql / MariaDB 10.3
PHP version:
7.4.3
Nextcloud version: (see Nextcloud admin page)
19.0.1
Updated from an older Nextcloud/ownCloud or fresh install:
From old NC, intially v15 or so...
Where did you install Nextcloud from:
ZIP -> no docker, native installation
List of activated apps:
e2ee: 1.5.2
Logs
Web server error log
2020/08/25 11:06:58 [error] 2054369#2054369: *244807 upstream timed out (110: Connection timed out) while reading response header from upstream, client: yyy, server: xxx, request: "PROPFIND /remote.php/webdav HTTP/1.1", upstream: "fastcgi://127.0.0.1:9000", host: "xxx"
Nextcloud log (data/nextcloud.log)
...
For me it looks like that MOVING the files from one nextcloud folder into the encrypted one caused the problem. As a DIFFERENT test user I could add new files to an encrypted folder without issues.
The text was updated successfully, but these errors were encountered: