Comments (5)
  • Add a Comment
  • Edit
  • More Actions v
  • Quarantine this Entry

1 DerekKeightley commented Permalink

Thanks for sharing. I have always wondered what the real implications of running no hd6 would be. The main fear I had was that it could he hard coded in places, I guess this could be one of those cases but good to know where it occurs. <div>&nbsp;</div> I tend to make it as small as possible (1 PP for eg) and assign designated paging space(s) elsewhere. Some will claim this is bad as paging spaces should all be equal, this may have been the case years ago but I have been assured this is no longer a concern.

2 AnthonyEnglish commented Permalink

I don't know whether the mksysb restore needed a paging space called hd6 or just any paging space, as long as it's in rootvg. It's good to get that observation about the equal size paging spaces. In the days when paging spaces were assigned to dedicated disks, you could understand the recommendation.

3 Tictac_NS commented Permalink

We leave hd6 in place, but in the last line of rc.local, we do a swapoff of /dev/hd6. It uses a little disk space, but it ensured that there was a paging space available when doing a mksysb restore to a system that does not have all of the "other" resources available. It also eliminates the problem of disparate paging space sizes on an active system. And lastly, it leaves us a quick "emergency" paging space allocation to buy us time if we get paging space utilization warnings, we can activate hd6 and take the small performance hit, while we look into what needs to be done to correct paging issues. <div>&nbsp;</div> Thanks for the article,... Ken C.

4 AnthonyEnglish commented Permalink

That's a clever approach. Good thinking Ken, and thanks for your comment.

5 RCeleste commented Permalink

Very useful , good to know. Thanks.