Hello Guys,
I need some help to find out if processes are running on remote server or not. I could do 'ssh' to do that but due to some security reasons, I need to avoid the ssh & get result from remote server.
Could you please suggest some that can be done without ssh or similar sort of utility ?
Greetings and welcome,
I don't understand why ssh ( which is what all admins want users to use for security...) is not allowed for security reason, how will you connect to remote server??
my understanding is you want to have a job running local on your remote host (using cron?) I suggest your cron job send an alert to syslog, there is on solaris and I think on other OS too a way of collecting the content of syslogs from remote servers ...
Bit vague requirement, processes must be running on servers 
Using, for instance, local5.info facility and logger command inside a shell script.
You configure that servers syslog to send logs to your central syslog server for that facility on IP using UDP or TCP.
Central syslog server will receive / parse etc. the logs messages.
Great choice for server side is any OS that can run rsyslog, which is good software, highly customizable.
Benefits arise when using ZFS filesystems on BSD/Solaris and Linux, when using compressed file systems to store logs.
Be sure to read about maximum message length limits and stuff, so you do not get into surprises.
Hope that helps
Regards
Peasant.
1) use remote procedure calls: you will have to write a program, it will be insecure and, frankly, i wouldn't want it in my data center, but it avoids ssh.
2) use telnet (or is this too similar to ssh?). Otherwise, see above but without the neet to write a program.
3) use SNMP. There is a reason why "SNMP" is thought to mean "security? Not my problem!", but it avoids the dubious ssh either.
4) create your own network protocol, then write a client/server-application which uses it to transmit the information. Report back in a few years to show us.
Bottom line: if you don't want to use secure means to connect to a sever you have to employ insecure means. Furthermore, if you want to avoid the obvious you will be restricted to the absurd.
Some more suggestions:
to Scrutinizer's reply i will add nagios could be a good option
Framed, and hanging in my office.
The remote server which I am talking about is highly important server so our security team doesn't allow any incoming connection (ssh or similar tools ) to this system, so need to avoid that.
But, after considering all other options, I have decided to do scp from remote server to local server & provide input to my script, I was trying to avoid this but for now, this is the easiest & security approved solution.
If at all you guys have any better idea, do share.
So how would you normally connect to the super-secure server? Console only perhaps?
It may well be just best to have a scheduled job on the server that posts the information somewhere. This could be by syslog (as mentioned before) or perhaps it could use ssh/scp to connect out and send them somewhere sensible. You could probably also get the server to HTTP-post a file to a webserver that has a cgi-script and just writes the file somewhere.
One wonders just how sensitive the server must be to not trust the administrator to sign on. You could open SSH incoming but limit access by coding up /etc/security/access.conf (or a file under /etc/security/access.conf.d) to restrict to an approved group. If you are paranoid, you could deny user/password access and force SSH keys to be used and make sure that the private keys matching the approved public keys are not shared.
There will be lots of ways to control SSH access, but we would need to understand the worries that the Security team have about the admin team and how far they would permit you to go.
Would a service account that has one function (to collect this information) be acceptable to them?
I hope that this helps with some options,
Robin