[GemStone-Smalltalk] Error with netldi, the connection to the repository is never made and the port becomes blocked

[James Foster via GemStone-Smalltalk]

Hi Ezequiel,

This is indeed interesting (and not in a good way). The netldid daemon process should simply start the new gem process and continue listening for other requests. With netstat you have confirmed that netldid is holding the socket. The next thing would be to get a C stack to see where netldid is hung.

With a clean startup before any connections, identify the PID for the netldid process (e.g., ‘gslist -l’) and send it a special signal:
$ kill -SIGUSR1 <PID>
The signal (which can be abbreviated as simply ‘usr1’) is interpreted by GemStone processes as an instruction to write the current C-level stack to the current log. Look for the netldi log (if you don’t know where it is try ‘gslist -x’) and go to the end and find the SIGUSR1 output. When idle, there should be something like ‘NetHibernateNoAction’ a few frames above main. 

Once you have the “normal” behavior, try your experiment that hangs and send it the signal. We are particularly interested in that stack. Wait ~30 seconds and repeat a couple times to see if it is in a tight loop or waiting on something. Let us know what that shows.

Also, I’d be interested in what MW thinks is its name and IP address, and if that name and IP can be pinged from ML. Try this in both the wired network and VPN configuration.

Finally, are there any other useful services on ML? For example, if it has a web server then it would be interesting to see how a web browser on MW reacted (though this is essentially what you confirmed with ping).

James


> On Mar 22, 2017, at 6:07 AM, brianstone via GemStone-Smalltalk <gemstone-smalltalk at lists.gemtalksystems.com> wrote:
> 
> Hi James,
> 
> First of all, thanks for helping me.
> 
> Ping and telnet (from MW) work with both connections direct and VPN
> 
> But if try now to login using Jade, then the the port becomes blocked or
> something,  because I cannot connect anymore with Telnet.
> 
> I've made another test:
> 
> I detected that is not only problem with my MW, because I tested connection
> from another pc with windows (connected using wire) and teh same error.
> 
> So, I started a session using vpn and one using direct connection, then I
> check in the server for current connections to the port using "netstat
> -putona | grep 50387".  50387 is the port of netldi.
> 
> I receive this result:
> 
> WITHOUT CONNECTIONS:
> 
> netstat -putona | grep 50387
> (Not all processes could be identified, non-owned process info
> will not be shown, you would have to be root to see it all.)
> tcp        0      0 :::50387                    :::*                       
> LISTEN      9362/netldid   off (0.00/0/0)
> 
> WITH ONE CONNECTION FROM VPN:
> 
> netstat -putona | grep 50387
> (Not all processes could be identified, non-owned process info
> will not be shown, you would have to be root to see it all.)
> tcp        0      0 :::50387                    :::*                       
> LISTEN      9362/netldid   off (0.00/0/0)
> tcp        0      0 ::ffff:133.94.1.30:50387    ::ffff:133.90.150.161:50546
> ESTABLISHED 9397/gem            keepalive (6975.87/0/0)
> 
> /*Note that a gem process is assigned to the connection*/
> 
> WITH ONE CONNECTION FROM VPN AND ONE CONNECTION WITH NETWORK WIRE:
> 
> netstat -putona | grep 50387
> (Not all processes could be identified, non-owned process info
> will not be shown, you would have to be root to see it all.)
> tcp        0      0 :::50387                    :::*                       
> LISTEN      9362/netldid   off (0.00/0/0)
> tcp        0      0 ::ffff:133.94.1.30:50387    ::ffff:133.90.150.161:50546
> ESTABLISHED 9397/gem            keepalive (6975.87/0/0)
> tcp      112      0 ::ffff:133.94.1.30:50387    ::ffff:133.94.1.11:54096   
> ESTABLISHED 9362/netldid  keepalive (7194.97/0/0)
> 
> /* Now,instead of a gem process, the netldi process is assigned to the
> connection from MW using network wire */
> 
> This is really strange. I have asked to network administrators and they said
> that there is no firewall in the network and that the problem may be my ML.
> 
> But I don't have any rule to reject connections in ML and the port is open
> too.
> 
> Thanks James.
> 
> Ezequiel
> 
> 
> 
> 
> --
> View this message in context: http://forum.world.st/Error-with-netldi-the-connection-to-the-repository-is-never-made-and-the-port-becomes-blocked-tp4939542p4939622.html
> Sent from the Gemstone/S mailing list archive at Nabble.com.
> _______________________________________________
> GemStone-Smalltalk mailing list
> GemStone-Smalltalk at lists.gemtalksystems.com
> http://lists.gemtalksystems.com/mailman/listinfo/gemstone-smalltalk