Hello, we use virtualhere with server (paid) installed on raspberry pies and client (either in service mode or not) on windows.
We are not 100% sure, but we think the time to unmount a device (triggered either via the UI or API) increased a lot since about one week or two, from about 6-8s to 24-25s. Same time from two different clients and two different servers.
Is it "normal" ? Can you assist?
.
You mean unmount a USB disk? Did anything else change in the last few weeks? Did you update the virtualhere server or client?
Unmounting speed
Yes, unmounting a device (smartcard reader). Actually I don't think anything changed as the two servers have different versions (4.6.5 and 4.7.2) and windows clients (5.5.7 and 5.7.6)
However, I just tested on one windows client an "auto mounting" (mount a local device with both client and server installed) and the unmounting is ok
.
Im not sure sorry. I dont think its a bug
Unmounting speed
You mean it is a normal behavior? can you at least help us investigate? Things we can try, logs to check...
This is a very important project (3 years of hard work in the health sector), we already bought 6 licenses and we intend to buy at least 15 and probably more.
Thank you
.
Sure, I can try to investigate more.
What exact smatcard reader is it? What is your latency?
To see latency, right click USB Servers -> About -> Statistics
.
Thank you
I tested 3 raspberry pies, the 3 of them in the same premise, 2 wifi and one ethernet (same switch, tcp latency <1ms)
The device is always the same : Gemalto USB SmartCard Reader : https://www.usmartcards.com/gemalto-desktop-readers/gemalto-id-bridge-ct30-reader
There are 8 readers per device.
Unmounting times (20-25s) are comparable either from my local computer or from our server (hosted in a same country datacenter: tcp latency 2-8ms)
Mounting time is about 1s
In the statistic view, I see all devices under 15ms except one with peaks up to 130ms
The latency is not affected by mounting and unmounting commands
Here is an extract of the vh_statistics file
09:54:13,Hub4,30859,104.53780409:54:14,CPx1,30771,0.46330009:54:14,Windows Hub,30796,0.29230009:54:14,Test hub,30858,2.47840009:54:16,Hub4,30859,106.43689709:54:17,CPx1,30771,0.40120009:54:17,Windows Hub,30796,0.36940009:54:17,Test hub,30858,7.34620009:54:19,Hub4,30859,108.14679709:54:20,CPx1,30771,0.49420009:54:20,Windows Hub,30796,0.48470009:54:20,Test hub,30858,5.69050009:54:22,Hub4,30859,116.27189609:54:23,CPx1,30771,0.42020009:54:23,Windows Hub,30796,0.24630009:54:23,Test hub,30858,7.30150009:54:25,Hub4,30859,77.70909909:54:26,CPx1,30771,0.47540009:54:26,Windows Hub,30796,0.31250009:54:26,Test hub,30858,10.06660009:54:28,Hub4,30859,116.52030209:54:29,CPx1,30771,0.38230009:54:29,Windows Hub,30796,0.32520009:54:29,Test hub,30858,5.69780009:54:31,Hub4,30859,106.91850309:54:32,CPx1,30771,0.49740009:54:32,Windows Hub,30796,0.36420009:54:32,Test hub,30858,5.86440009:54:34,Hub4,30859,107.36239609:54:35,CPx1,30771,0.47560009:54:35,Windows Hub,30796,0.23290009:54:35,Test hub,30858,3.34120009:54:37,Hub4,30859,3.89930009:54:38,CPx1,30771,0.53480009:54:38,Windows Hub,30796,0.39500009:54:38,Test hub,30858,6.26270009:54:40,Hub4,30859,2.60970009:54:41,CPx1,30771,2.06630009:54:41,Windows Hub,30796,0.42850009:54:41,Test hub,30858,5.47070009:54:43,Hub4,30859,116.64720209:54:44,CPx1,30771,0.48960009:54:44,Windows Hub,30796,0.32220009:54:44,Test hub,30858,14.91670009:54:46,Hub4,30859,31.61179909:54:47,CPx1,30771,15.09830009:54:47,Windows Hub,30796,0.32810009:54:47,Test hub,30858,15.56940009:54:49,Hub4,30859,72.08329809:54:50,CPx1,30771,0.40060009:54:50,Windows Hub,30796,0.351500.
Hub4 has high latency. Are all the servers basically similar times for unmounting via the client?
I assume you are running the latest drivers for the CT https://supportportal.gemalto.com/csm?id=kb_article_view&sys_kb_id=0adc96844f350700873b69d18110c76a&sysparm_article=KB0016522 (they seem quite old anyway)
.
Hello again,
Good news! After bouncing from one inconsistency to another, we finally traced the long unmount time to something completely unrelated to the readers or drivers.
Recently, someone changed the hostname of each Raspberry Pi but forgot to add the new name to
/etc/hosts.Because the name could not be resolved locally, every unmount triggered a 20-second DNS fallback/timeout.
After adding the correct entry (or reverting the hostname), unmount time is back to around 4 seconds.
So VirtualHere itself was fine — it was the missing host file entry that caused the slowdown.
Thanks a lot for your assistance anyway.
I hope the note above helps if another user runs into the same issue.
.
OK thats great news, thanks for letting me know!