HP-UX

I don't know if this is better suited for the application section, but here goes.

We are currently running HP-UX 11 as our database server. The database is Progress version 9.1C.

As of late, some of our batch processes that run on the UNIX db server are erroring out because of what appear to be memory issues(at least according to Progress). The db error messages indicate that either there are too may subprocesses and it cannot fork, there is not enough memory to execute the request, or there is not enough memory to allocate a sort buffer. There does not appear to be a consistent pattern to when this happens. Also, we don't appear to be exceeding swap space.

Not sure what else it could be. If any of you have encountered similiar problems, how did you resolve it? Any and all feedback is appreciated.

-Ed

First, this is probably the best place for the thread. The symptoms you describe lead me to suspect swap space despite your comments to the contrary. Run "swapinfo -t" and look at the last line to be sure. Also consider getting the glance program. It is a product, but it doesn't cost much and it can diagnose what the problem is. There may be a free trial version you can get.

You said it wasn't a swap space issue, but as Perderabo suggested take another close look at that, and if you have more than one swap space, there's no real guarantee which one will be selected at any point in time.

The first issue, can't fork, almost sounds like you need to do some kernel tweeking. The database owners should be able to give you some ballpark kernel settings, however thats all dependant of DB size, how much/how often updates, queries take place for your installation; however they should be able to give you some min. settings. If you've not already tweeked these, max_thread_count might be a place to look, however there are many other paramaters worth looking at. Doing in a production envoriment is not ideal. Definately document well, what your original setting were, and exactly what was changed. Best done with someone with some experience with kernel tuning.

You may have errors in syslog.
If you have glance, it can give a very good picture of what is going on on your system.
If you have hardware errors, you may have emails to root.