What to do short of a rebuild for RME 3.5/IDU 13 on Solaris?

[getwithrob]

I have 3- RME servers @ IDU 13 on Solaris all the way patched. The servers are at or above capacity but work fine other than this issue:

On a VERY regular basis 3-4 times/week when trying to do anything where I have to click on the All Devices symbol, I get the message No Devices to Display. All I can do when this happens is do a dmgtd stop and start.

Is there anything I can do short of rebuilding that could help me out so I don't see this ugly symptom so very many times each week?

Thanks for input, rob

[Joe Clarke]

You are most likely hitting a known deadlock bug in RME 3.5 which will never be fixed. The workaround is to restart DIServer, IcServer, and JRunProxyServer. You can do these all on one line:

pdterm DIServer IcServer JRunProxyServer

pdexec DIServer IcServer JRunProxyServer

All users will need to log back into to CiscoWorks.

If that does not work, you may be hitting CSCin35199 which will require you to restart all daemons before LMS will work again. Neither of these problems affect LMS 2.5 and higher.

[getwithrob]

Thanks for the input.

What I find interesting is this started happening about 1 year ago. The 2 years prior to that, I seldom ever had to bounce the app. Is there perhaps DB inconsistencies that could be corrected short of a rebuild?

[Joe Clarke · Accepted solution]

Doubtful. If it's the former problem, the issue is periodic inventory collection jobs lock the database tables for exclusive access, and the GUI cannot access them to display the devices. If it's the latter problem, it is purely a servlet memory leak. Besides a database corruption issue would not be resolved by a simple restart of the daemons.

[getwithrob]

Thanks for your valuable and always accurate input!