FB 5.0.1: error at the end of the backup process
Server mode: Classic or Super (tested for both) Max parallel workers: 64 Parallel workers: 16
At the end of database backup gets the following message:
12:42:46,47 C:\Program Files\FB5>gbak -b "G:\XXXXXX.FDB" "G:\XXXXXXX.bk" -pas XXXXXXX -user sysdba -se localhost/3056:service_mgr
gbak: ERROR:Error reading data from the connection.
gbak:Exiting before completion due to errors
There is a record in the firebird.log:
XXXXXXXX Wed Apr 3 12:57:23 2024
INET/inet_error: read errno = 10054, server host = localhost, address = ::1/3056
Visually, the resulting bk file looks fine, though I didn't check how the database will be restored from it.
Title shows "restore", command line shows "-b", i.e. backup. What is correct ? ;) How it can be reproduced ? Or you have a crash dump ?
Sorry, fixed the title. This is an issue with back up now. I'm still in my thoughts about that restoring error, that was fixed recently.
I need more info to investigate the issue. Tried to reproduce with db that I've used to reproduce #7942 - no luck.
Any new info, test case probably ?
I will wait until this commit goes into the FB 5 build and then recheck.
https://github.com/FirebirdSQL/firebird/commit/e96fc77d504740d2d8e7b69d9594c2700877c9a3
Hmm... that was about restore. Anyway, it was fixed in v5 first, two days ago.
Look like 5.0.1 will not include fix for this "Elusive Joe". Should I close the issue as not reproducible ?
it is reproducible 100%. I'm just weighting procs and cons of sending sensitive data.
It is up to you, but I can't fix something that is not reproducible and with no info.
@hvlad sent instructions how to get the db by email
Received your email, thanks
Still can't reproduce it
I'm closing it. Let's assume it is somehow related to the particular configuration of Windows 2019 Server and three Firebird servers (2.5, 3.0, 5.0) running on different ports.