QFieldSync takes nearly 5 minutes on a network drive but only 14 seconds locally

Ask the Community
1

Severe QFieldSync delay observed when the QGIS project is stored on a mapped network drive.

A test was done on our end to compare the difference using the same project and the same sync/apply operation:

Network drive: 4 minutes 47 seconds
Local C: drive: 13.73 seconds

The network-drive test was approximately 21 times slower. During the delay, QGIS appears to be frozen, and the synced changes do not appear on screen until the process finishes.

Environment:

QGIS 3.40.15-Bratislava
QFieldSync 4.22.3
Windows 11 Pro
15.5 GB RAM
Mapped Microsoft Windows network drive

The result strongly suggests that the project’s storage location affects QFieldSync performance. Is this a known issue when using QFieldSync with projects stored on network drives? Are there any recommended settings or configuration changes that could reduce the delay while keeping the project and data on the network drive?

image
image1050Ă—758 122 KB

Steps to reproduce:

  • Open a QGIS project stored on a mapped network drive.
  • Open the QFieldSync plugin.
  • Apply the synchronized changes to the project.
  • Measure the time required for the process to complete.
  • Copy the same project and data to the local C: drive.
  • Repeat the same QFieldSync operation and compare the processing time.
2

silly question but, what are the data rate differences when reading and writing files between those two options using other methods? In other words, how did you discarded that this is not an issue with the network drive (or the sharing method, network, etc…)?

3

Hello: he cloud drive I mentioned physically sits in the office and is accessed company wide through intranet. And access to the internet is business fibre. I don’t think there’s speed issues here.

4

It sounds pretty close to the difference in write speeds between a big, slow hard drive for storage and a nvme ssd typically found on system drives nowadays. Add any possible network congestions, protocol overheads and whatnot and it doesn’t sound too implausible. That’s why I asked if you measured the speed differences between those two data storages any other way.

I’ll see if I can reproduce this issue on the weekend and try to report back.

5

we have now considered data speeds being possible and have switched the test machine from wifi to ethernet. Anecdotally, sync times have appeared to drop within the 2-5 minute mark, but remain behind the local speed test of 13s