I have a client program that communicates via a local Windows service to a server process. These three programs are implemented by ourselves. To analyze an error situation, I'd like to take crash-dumps of all three processes at once, so that there are no inconsistencies between them. The trigger is an exception in the client.
In the testing environment, the server runs on the same machine. In that case I can attach to the processes in question using .tlist and .attach. Once the exception is thrown, all processes get suspended and I can do a .dump on each single dump.
Is it possible to achieve the same result if the server is running on a different machine? How?
I have tried: attaching WinDbg to the serverside process, then running a remote debugging server (.server), connecting to it from a new WinDbg instance on the client. But then I can't attach to local processes, since the .tlist and .attach command operate on the server instead of the client.
Related
There is a powershell script on a windows server(2016) that runs basic network connectivity test such as ICMP ping Test-Connection/Test-NetConnection cmdlets against couple of windows servers ~3k. The results of those commands are streamed as events to Splunk http event collector. The target windows servers are located within retail stores spread across the country. The event data which is streamed to Splunk has details such as the hostname, status(online/offline), center-number(basically the retail store identifier). The powershell script is scheduled to run every ten minutes. However it taking a long time to complete a single run. As new stores are opened, this is only going to delay the execution of the script. Now it takes ~20 min. to ping 3k servers.
Can Ansible be leveraged to trigger the powershell script by passing the store/center identifier as a parameter? Due to security restrictions, the in-store servers cannot receive ICMP ping/echo request from any machine. The windows server hosting the powershell script is the only machine which can successfully run ICMP ping commands against the server. With ansible, probably we are thinking that the remote powershell script can be launched in a for loop(of all stores) as a async task perhaps.
The powershell script can run a quick server availability test for a store, stream the result and then quit which is more efficient than having it run for nearly 25 min. The only concern is ansible will be making multiple connections to remote WinRM service on a single windows host. The host has 16 Gb of RAM.
Is this a good approach?
I use Delphi. I have a backup application which has to run constantly on my customers machines. Communicaton is done via TCP/IP Sockets & data is transferred via FTP.
My problem is that sometimes when the customer upgrades his AntiVirus or I send a upgrade the Firewall blocks my Application even though setup was done via elevated rights. How do I overcome this. I had thought running my communication portion from a Windows Service Application would solve the matter.
I am Writing a Datagram Client and Server using eclipse with the class DatagramSocket. My question is: first I run the server from Eclipse, but when I run the client the server will continue work or just terminate ?
I wrote a server discovery system while attending computer networks course at my university, have a look:
https://github.com/colonder/Rudolph-server-discovery-system
It was a bit ago but as far as I remember both server and client are separate threads so one shouldn't kill the other, but I'm not so sure about running both things in the same Eclipse instance. When I was coding it at home I had only one machine available, so I guess it's possible to run both of them in the same Eclipse instance.
I am Unix java developer trying to help a datastage developer, so out of my aquarium.
The datastage process connects to a database hosting financial transactions on a unix server. there is a datastage process for migrating financial transactions to the ACCOUNTING system. The ETL developers for one reason or another have specified they cannot run one or more specific ETL while in-taking new financial transactions and have specified the process that inserts transactions into the DB be stopped.
me java geek thinks have some process checking a service running at port 55555 would be perfect. But we cannot find a way for datastage to create a socket connection to a port to check. I don't do datastage so I don't know how to work around it's limitations.
The ETL developer thinks a cron script running every minute that inserts an up/down status for the process into a special table would be perfect. I think it is a waste of cpu.
I cannot be the only company that cannot run an ETL when some process is running on a remote system.
How did you solve this issue? Is there a way to connect to a remote servers socket and run the service from datastage???
thanks
after a bunch of discussion.
options we found
Add a step to the start stop server scripts that writes process status to a table. pro: easy to implement. con: not turely accurate (some geek like me is likely to bypass the start/stop script and run only build/run the executable bypassing the start stop script and bypassing the step that inserts the status.) No network and InfoSec paperwork
Cron based script that updates the table with the status on a minute by minute basis. what a pain!!!! No network and InfoSec paperwork.
A script made available to the network through inet or xinet. Problem is datastage ETL developer does not know how to connect to a socket via C or java program. Creates Infosec and network paperwork issues.
New web service (there is a tomcat server serving up a number of web services) Problem is datastage ETL developer does not know how to connect to a socket via C or java program. Creates Infosec and network paperwork issues.
options 3 and 4 are accurate and realtime. options 1 and 2 opens up the possibility of inaccuracies by bypassing process, but that opens up a different can of worms.
We are probably going to implement option 1
I am using a perl script to log onto a remote router running JUNOS and get few information through the NETCONF port.
I am using the following library in order to do this
Net::Netconf::Manager
Net::Netconf::Device
Net::Netconf::Access
The script runs on the routers present on the same network but is failing to log into the routers that are on different networks, where as i am able to log onto the routers manually through Netconf over ssh.
The arguments I am parsing looks fine. Could any one help me resolve why this is happening.
If I need to do any checks in configuration (configuration looks fine to me) prior to running the script.