10.5 stops polling proactive script

[splurzgalonk]

I have 10.5 installed on a Mac and three proactive scripts with a dozen or so sources in each. The backup runs fine until a certain point in time (seemingly erratic), when the polling simply stops. Reboot or a bit simpler, disabling and re-enabling of the retrospect engine makes the polling work again.

 

What's happening here? Any idea on how to prevent this "sleep mode" from happening?

 

cheers,

Roland

[Lennart_T]

Does it actually stop polling? Or just pauses a long time?

Or it doesn't find the clients?

Or it doesn't find the media sets?

 

Any messages in the proactive window?

[splurzgalonk]

It stops seeing any client that appears in the network, even if they are available for hours over gigabit ethernet and I was able to check them with ARD. If the clients are not found, I usually get a message saying so. Not in this case, the engine simply doesn't see any clients appear. Media sets are fine too. No errors, no sign of problems. The scripts simply sit there, all green, and they do NOTHING at all, as far as I can see.

 

Restarting the engine immediately picks up the clients and starts backing up again. It really just seems to forget to listen. 

[Lennart_T]

I only run proactive in Windows, not Mac OS.

 

Here is what it looks like:

"Busy" means the media set is busy (with running a backup)

"Source" means the source can't be found.

"Executing" means that client is being backed up right now.

"Media" (not shown) means the media set can't be found.

 

How does it look on your server when it has "stopped polling"?

[splurzgalonk]

First of all, thanks, Lennart.
I don't have any idea, where I would find the screen you've posted. Seem like the windows UI is quite different to the Mac. 

 

I derive the fact that it stopped polling by simply looking at when the last backup of ANY client on the net happened. And after a certain point there is no backup made of any client.

As soon as I restart the engine, the script picks up those clients visible on the net and backs up properly. 

 

I'll have to wait until it happens again to see what the states of all the sources is at that moment. As far as I can remember, the state was "ok" for all (= green icon in Mac UI).

[dkwebber]

This seems to be exactly what is happening with my proactive scripts as well. The proactive scripts seem to launch as scheduled first thing in the morning (several launch at 8:30 am, others launch at 9:00 am), the different proactive scripts will back up several clients as expected, then the scripts just seem to "stop" ; they do - not - seem to be polling or running after about 10:30 am.

If I stop and then re-start the engine, several proactive scripts will immediately launch and begin the backup of another client ... for a time. Then, for no apparent, the scripts just cease to poll or run again. That is, until I restart the engine. Again.

So, it's not just my set-up here, apparently. As "Roland" stated above (from his post) ... "What's happening here?"
This was not the case before I upgraded to Retrospect 10.5

 

I have a current support ticket with Retrospect Support regarding this very issue. So far, Retrospect Support has not been very helpful. The most current response I had from Support there was that the problem could be caused by some type of corruption to the config80.dat file. Replacing my config80.dat is certainly the nuclear option, but I did not have this problem before 10.5 upgrade. Does anyone have another option? Re-installing and re-configuring 24 scripts with 185 clients and 15 servers and volumes would be a week of hard work.

 

Any insight appreciated,

Don

[twickland]

We had a problem back in Retrospect 8.2 that was solved by uninstalling and reinstalling the console and engine. While that was a few generations ago, it may be worth a try.

 

Have you noticed any pattern as to when the proactive script stops, like after attempting to connect with a particular client or subnet?

 

Have you tried creating a new test script to see if it misbehaves the same way?

 

You might also try dragging Config80.dat to the desktop (quit the console and stop the engine first) and creating a clean test setup with some subset of clients and a new proactive script to see if the behavior improves or if the problem recurs when a particular client is added to the script.

[Glibbs]

Same problem here.

 

Have done all that's been recommended:-

 

Re-done the config file several times.

 

Re-installed the whole App after erasing everything retrospect.

 

Haven't re-initialised the disk.

[splurzgalonk]

I still have that issue and I can't really pinpoint it to when or why it happens. Every now and then, the proactive scripts simply stop polling. The only way I can make it run again is to reboot the system. In the background, I see that retrospect engine is constantly running 100% (1 core) on a multi core machine.

 

Another retrospect server of mine repeatedly looses connection to an erratic amount of machines that are on the net visible and accessible. I have to use the "Locate" button to "reactivate" that very connection and as soon as I do that, the proactive script, which runs and backs up other sources on the same net during this issue, starts to see those "Located" machines again.