We are using Samba 2.2.8 on Redhat Linux 2.1.
For the past few months, we've been getting the following errors:
[2005/09/16 12:22:28, 0] lib/util_sock.c:read_data(436)
read_data: read failure for 4. Error = Connection timed out
This is on different PC's at different locations.
One site was even running SCO Unix (on the same version of Samba).
The PC's run Windows XP (most on Service Pack 1, some Service Pack 2).
We have approximately 150 locations and only about 20 have reported this problem
(some more than once).
Typically the manager at the site will be in email when the PC will
"hang". Sometimes they are saving documents in Word, sometimes
working on our Intranet.
They will boot the PC and it will come up but not appear to be on the LAN/WAN.
By the time we at 2nd level support get called we can connect to the PC (through
PC-Anywhere) and even see the Samba shares from the PC.
We boot the PC (through PC-Anywhere) and it comes up normal.
The manager's tell us they may have booted several times before they call us
and sometimes the PC is normal after the first boot, sometimes it does not come
back normal.
I know we should upgrade Samba, but beyond that we are curious as to why this
just started to happen (also why it's not happening everywhere since all our
sites are essentially exactly the same).
Given the problem, we're not even sure Samba is a "culprint" or a
"victim".
Any tips at what to start looking at would be appreciated.
