Re: bin/56129 (ftp(1) should keep control session duringlargexfer)

[Izumi Tsutsui <tsutsui%ceres.dti.ne.jp@localhost>]

The following reply was made to PR bin/56129; it has been noted by GNATS.

From: Izumi Tsutsui <tsutsui%ceres.dti.ne.jp@localhost>
To: rvp%SDF.ORG@localhost
Cc: gnats-bugs%netbsd.org@localhost, lukem%netbsd.org@localhost, tsutsui%ceres.dti.ne.jp@localhost
Subject: Re: bin/56129 (ftp(1) should keep control session duringlargexfer)
Date: Thu, 24 Jun 2021 00:21:27 +0900

 > > Today I cannot reproduce the control session timeout on
 > > my HP9000/425t (68040 25MHz / 48MB):
 > >  https://pbs.twimg.com/media/E4f1dL1VkAc3hEh.jpg
 > > either on GENERIC and /usr/bin/ftp or RAMDISK kernel crunched one.
 > >
 > > I'll also check again slower 68030 machines (luna68k and x68k).
 > >
 > 
 > From the screenshot:
 > 
 > 41046620 bytes received in 06:00 (111.23 KiB/s)
 > 
 > That's just 6 mins. Try downloading `debug.tar.xz' which should
 > take more than 15 mins. From your past screenshots, my feeling
 > is that any _single_ transfer which leaves the control connection
 > idle for >15min.
 
 The similar timeout occured 6minsq or 8 mins so there might be
 more conditions, but anyway I'll try reproduce on slower machines
 or large files.
  https://pbs.twimg.com/media/E3uicvFUUAYQ85j.jpg
  https://pbs.twimg.com/media/E3uspBhUYAIFURZ.jpg
 
 > will simply be dropped silently by your FTP NAT.
 
 Yes, there are two NAT (NetBSD/i386 9.2 npf with ipfilter compat one,
 and NTT(the line service provider)'s "Home Gateway"), so they could be
 problematic.
 
 Thanks,
 ---
 Izumi Tsutsui