Topic: Cable (with SME 5.6)

[Luke Drumm]

Howdy,

I've got Telstra cable , and as such, am at the mercy of the stupid heartbeat system.

Unfortunately my logs reveal that bpalogin is able to login but not 'hear' the heartbeats. eg:

Feb  9 05:23:39 delta bpalogin[12072]: Logged on as ******** - successful at Sun Feb  9 05:23:39 2003
Feb  9 05:30:39 delta bpalogin[12072]: Timed out waiting for heartbeat - logging on

The bpalogin's bootup message is as follows:

Feb  8 12:23:09 delta bpalogin[2263]: Auth host = dce-server.nsw.bigpond.net.au:5050
Feb  8 12:23:09 delta bpalogin[2263]: Listening on port 5050

Charlie's e-smith-bpalogin rpm has been installed along with version 2 of the bpalogin software (although I've tried the same thing with the old reliable version 1.3 and received the same result).

Here's the customisations I've done:

/etc/e-smith/templates-custom/etc/bpalogin.conf/20username
   username **********

/etc/e-smith/templates-custom/etc/bpalogin.conf/30password
   password **********

/etc/e-smith/templates-custom/etc/bpalogin.conf/40authserver
   authserver dce-server

/etc/e-smith/templates-custom/etc/bpalogin.conf/60localport
   localport 5050

/etc/e-smith/templates-custom/etc/bpalogin.conf/70authdomain
   authdomain nsw.bigpond.net.au
   
/etc/e-smith/templates-custom/etc/rc.d/init.d/masq/45AllowBPALogin
   /sbin/iptables --insert INPUT 9 --protocol udp --source 61.9.192.13 --destination $OUTERNET --dport 5050 --in-interface $OUTERIF --jump ACCEPT
   {
       $OUT = allow_tcp_in(5050,1);
   }


Any ideas on what I've missed?

[Kelvin]

Hi again Luke,

I too had this problem and was in discussion with Damien Curtain about it. However, I found that the messages stopped after about half an hour.

I am not using Charlie's contrib, just bpalogin (version 2.0-1) straight from Sourceforge's site. My SME is version 5.6.

The main changes to the bpalogin.conf file (aside from username and password ) are :-

authserver spr3

authdomain nsw.bigpond.net.au

localport 5050

minheartbeatinterval 60

When I first restarted bpalogin, I was still getting the heartbeat timeout every 7 minutes. But after about half an hour, the messages stopped and it's been going OK since (fingers crossed). My thanks to Damien for his iptables and other  settings.

Hope this helps.

Kelvin

[robert]

Hi There
do you wat to try the Howto at:
http://www.synapticserver.com/bpalogin_2howto.html

We haven't tested it on 5.6 so your input would be valued!Kelvin wrote:
>
> Hi again Luke,
>
> I too had this problem and was in discussion with Damien
> Curtain about it. However, I found that the messages stopped
> after about half an hour.
>
> I am not using Charlie's contrib, just bpalogin (version
> 2.0-1) straight from Sourceforge's site. My SME is version 5.6.
>
> The main changes to the bpalogin.conf file (aside from
> username and password ) are :-
>
> authserver spr3
>
> authdomain nsw.bigpond.net.au
>
> localport 5050
>
> minheartbeatinterval 60
>
> When I first restarted bpalogin, I was still getting the
> heartbeat timeout every 7 minutes. But after about half an
> hour, the messages stopped and it's been going OK since
> (fingers crossed). My thanks to Damien for his iptables and
> other  settings.
>
> Hope this helps.
>
> Kelvin

[Kelvin]

Hi Robert,

Yes, I used that contrib in my last server (5.1.2). I would like to see if version 2.x of bpalogin will work properly or not as it has new features I would like to use. Also version 1.3 has a problem recovering from an extended outage -- the server keeps trying to logon and keeps failing even though Bigpond is back up and running already. Only a server restart fixes the problem. I am hoping this problem has been fixed in version 2.x.

I know version 2.x did not work properly under previous SME versions -- I have tried it on 5.1.2 and it keeps getting the heartbeat timeout issues. Under 5.6, at the moment, heartbeat timeouts still occur, but nowhere near as bad as under 5.1.2. However, I'll keep monitoring it as I need the connectivity alive or it will impact on my work.

Kelvin

[robert]

Kelvin wrote:
>
> Hi Robert,
>
> Yes, I used that contrib in my last server (5.1.2). I would
> like to see if version 2.x of bpalogin will work properly or
> not as it has new features I would like to use. Also version
> 1.3 has a problem recovering from an extended outage -- the
> server keeps trying to logon and keeps failing even though
> Bigpond is back up and running already. Only a server restart
> fixes the problem. I am hoping this problem has been fixed in
> version 2.x.

Upgrading to verison 2.x seems to have no effect.

As a workaround we just type
service bpalogin restart
on the command line. This gets it to login again. The error  messages about failing to logon no heartbeat continue. I did hear that the time-out on the heatbeat had been changed . Do you know anything of this?

 
> I know version 2.x did not work properly under previous SME
> versions -- I have tried it on 5.1.2 and it keeps getting the
> heartbeat timeout issues. Under 5.6, at the moment, heartbeat
> timeouts still occur, but nowhere near as bad as under 5.1.2.
> However, I'll keep monitoring it as I need the connectivity
> alive or it will impact on my work.


> Kelvin

[Kelvin]

Hi Robert,

>The error messages about failing to logon no heartbeat continue. I did hear that
>the time-out on the heatbeat had been changed . Do you know anything of this?

On my server, the time out issue happens for about half an hour after initially installing and starting it up. After that it goes away. It still surfaces once or twice every day but only for about half an hour to an hour each time then goes away again. I too hear a rumour (not confirmed -- hence a rumour !) that the hearteat timing has changed. I also saw at sourceforge a post from some guy who was looking at the source code of the program and thinks he has found a logic bug in the heartbeat detection routine. Maybe it's a combination of factors.

>Upgrading to verison 2.x seems to have no effect.
>
>As a workaround we just type
>service bpalogin restart

Ho ! Ho ! Now that's a worry ! I guess I'll wait till the next serious outage and see how badly it handles / does not handle it. Perhaps, I'll try and find / write a script to check if bpalogin managed to relogin and if it has been failing continuously for a given time, automatically run the restart command. Anyone care to write one for me

Kelvin

[robert]

> On my server, the time out issue happens for about half an
> hour after initially installing and starting it up. After
> that it goes away. It still surfaces once or twice every day
> but only for about half an hour to an hour each time then
> goes away again. I too hear a rumour (not confirmed -- hence
> a rumour !) that the hearteat timing has changed. I also saw
> at sourceforge a post from some guy who was looking at the
> source code of the program and thinks he has found a logic
> bug in the heartbeat detection routine. Maybe it's a
> combination of factors.

> >service bpalogin restart

Well it beats restarting the server...

Thats not such a bad idea. If I have a moment
 I will write it!
> Kelvin