Glance Sys Mem Reporting 16.0tb

[gstonian]

One of our HPUX 11.23 Itanium Servers are reporting a huge amount of Sys Mem

Upon rebooting it recorded using 1.5gb as per below. After starting up our Application Systems (inc. Oracle) glance then reported Sys Mem of 16.0tb

Before
Total VM : 2.2gb Sys Mem : 1.5gb User Mem: 1.2gb Phys Mem : 16.0gb
Active VM: 2.0gb Buf Cache: 1.0gb Free Mem: 12.3gb FileCache: na

After
Total VM : 33.6gb Sys Mem : 16.0tb User Mem: 14.6gb Phys Mem : 16.0gb
Active VM: 6.9gb Buf Cache: 1.3gb Free Mem: 420mb FileCache: na

Given it looks related to the application, whats the best way to investigate this and prove what's happening.

I can get the applications shut down but doesn't really resolve the issue given these apps have been happily running on this server for last 5+ years & only seems to have happened recently.

Any advice welcomed & points awarded as always :)

[Dennis Handly · Accepted solution]

What does "swapinfo -tam" show?
Have you installed any glance patches? Sys Mem: 16.0tb is a little too much.

See these threads with Ntb being reported:
http://h30499.www3.hp.com/t5/System-Administration/Mem-100-full/m-p/5236621#M469188

http://h30499.www3.hp.com/t5/General/RAM-is-only-used-at-40-but-swap-is-already-kick-in/m-p/4775148

The latter mentions: PHKL_41972

>these apps have been happily running on this server for last 5+ years & only seems to have happened recently.

It seems that glance is returning bogus info, especially if everyone is happy but it.

[Dennis Handly]

>ME: See these threads with Ntb being reported:

Both of them mentioned 11.31, not your 11.23.