Child-Cache-Problem (Timeout while transfering data)

From: <>
Date: Tue, 13 Oct 1998 12:33:58 +0100


I have a child-proxy which gives it's users timout-errors, while it is getting
data from its parent. Mainly this occurs while downloading big ftp-files

The configuration:

Parent-Cache <--64k/s--> Child-Cache
1.NOVM.2[02] 1.NOVM.22 or 1.1.6

The Child-Cache requests a ftp-object which is already in the cache of the

It asks per ICP for it an gets (as expected) an UDP_HIT

907162812.057 10 UDP_HIT/000 58 ICP_QUERY - NONE/- -

Then it starts to download the object... I can see the transfer on my
networkmonitor. It runs at maximum rate (all available bandwith on the

But (and that is the problem) the requesting user sees nothing, his browser
does not show that there is data approaching... after 900 seconds it gives a
READ TIMEOUT to the user.

Child: 907163978.150 900242 ERR_READ_TIMEOUT/400 801 GET - PARENT_HIT/
Parent: 907163905.430 1092862 TCP_HIT/200 4522406 GET - NONE/-

(ok, the clock of the Child was 5min out of sync)

on smaller objects the Browser also stalls. After _completing_ the transfer
between the caches it delivers the data as fast as it can to the client.

An upgrade from 1.1.6 to 1.NOVM.22 does not change the behaviour.

If I let the child go directly, there are no problems... it loads the data
from the source and delivers the data simultanously to the client.

How can i get this behaviour also with a proxy-cascade? (other proxy-chains
here work without problems...)

Both proxies are configured to the installation-defaults.


Systemadministration            Email
Abacom-Internet GmbH & Co       WWW
Konrad-Zuse-Str. 1              Tel                05271/967-0
37671 Hoexter                   Fax              05271/967-267
Received on Tue Oct 13 1998 - 05:39:21 MDT

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:42:28 MST