this is not my case as far as i can see, i have no iptables rules at all and no firewall in the middle to block.
i tried to increase limitation of xinet.d and this does not change a thing.

anyone else can think of something? anything? i will try all...


Regards,

Adi Spivak

IT Manager

 

17 HaYetzira St, "Mordot HaCarmel" Industrial Park, Yokneam Elite

Mail: P.O Box 687, Yokneam 20692. Israel

TelFax.: (+972) 48-290-124. (+972) 48-290-100

www.fit-pc.com

www.compulab.co.il


On 10/05/2011 09:15 PM, Debra Baddorf wrote:
My network guy made the change, but if was in some priv settings on the server -- allowing the
client to initiate a new connection.   Totally outside of amanda.  
I think we were running free-BSD at the time.  I can't find similar files right now under Linux,
but it might have been in rules.ipf   or something like that?   It depends on what connection
security you have turned on, I guess.

But I hope it gives you someplace to explore ..... if nobody has any internal-to-amanda fixes for you.
Deb


On Oct 5, 2011, at 1:54 PM, Adi Spivak wrote:

Can you alaborate on what you did to fix this? What permission is missing? What did you change? The client or the server?
--
Thanks,
Adi Spivak

Debra Baddorf <[email protected]> wrote:
One problem I had early on,  which you might be having:
When amanda starts a backup,  the connection is held open for XX amount of time.
The client can reply back on that same connection ..... unless it takes longer than that
to collect its data and get started.   That could change as the data increases in size.
If the initial connection is no longer open,  the client has to start a NEW connection
back to the server.
  On my machines, that required a different level of privilege than using a connection
that the server had started.  My client didn't have connection to initiate a connection.
Took us a while to find and fix that....

Deb Baddorf
Fermilab


On Oct 5, 2011, at 2:41 AM, Adi Spivak wrote:

hi.
Amanda version 2.6.1p1

this server has worked perfectly for more then a year.
now, i get ACK error all the times, mostly on random servers/workstations, however it is consistence in self error, meaning the server itself gets ACK error on itself all the time.
nothing has been changed as far as i know (the main configuration was not changed) as i did not change anything until the error started to show up.
then i followed the guide here:
http://wiki.zmanda.com/index.php/Selfcheck_request_failed
and tried everything ther (i only use bsdtcp amservice returns ok on all of them,checked firewalls and xinetd configuration, tried upgrading the clients on windows machines and even started to use the hosts file) but nothing works.

as i understand from the guide since amservice checks out ok, this is a server issue, but i could find none.

i also tried increasing etimeout, dtimeout and ctimeout.
i even tried weird stuff i found like network settings:
/sbin/ifconfig eth0 txqueuelen 10000
/bin/echo 180 > /proc/sys/net/ipv4/tcp_keepalive_time
and eventually i even started to replace all my backbone switches to check if this is my network issue (highly unlikley as it fails on itself too, but i am running out of options).

anyone knows what else to check? please do not advise an upgrade as i currently does not have a backup for my backup server and not in a time for buying new hardware or even have much time for tests to see what will break if i upgrade.

thank you.
--
Regards,
Adi Spivak
IT Manager
<CompuLab-Logo-200-cut-out.png> 

17 HaYetzira St, "Mordot HaCarmel" Industrial Park, Yokneam Elite

Mail: P.O Box 687, Yokneam 20692. Israel

TelFax.: (+972) 48-290-124. (+972) 48-290-100



Reply via email to