Re: HP9000+LVM+Online

Article: 13082 of comp.databases.informix >> >> Andy >> >> I got the impression that Mike's system only had one chunk (therefore >> one page cleaner) which was striped underneath using LVM. > >Put one page cleaner on a five-disk system and you've already got your >checkpoint performance down to 20% of what it might have been. Agreed. The more page cleaners the better. >> In general, this would (theoretically) quadruple read throughput i.e. >> given a long random list of read requests, it would get through them >> four times quicker. > >Sorry, I just don't see how you arrive at that. AIUI a well-distributed >chunk design can do anything you can do with LVM. It's the same idea as v6 split reads except it's over 4 drives. Getting a well distributed chunk design is sometimes difficult when you have a lot of small tables. And if you have a heavily updated small table, using LVM is easier to use. >> For writes, the application inserts 250,000 times (20%) and updates 1M+ >> times (80%). Again, LVM would mean that updates are quicker because >> reading the record + index leafs would be faster. > >Nope. Don't see why. At least not over and above what you could do with >standard OnLine facilities. Assuming the row size is 200 bytes, this table would be 50MB-ish. On pre-V7 this would be a pain to distribute evenly over multiple disks so I would leave this table on one chunk. Using LVM, I can (theoretically) quadruple the update throughput of this table. I would then have to compare this gain against the page cleaning throughput loss. If this table is small and on one chunk, there is no page cleaning loss. >> For sorted writes, you will still get a sorted write on each of the >> ............... As you can see, the actual physical write to >> each disk is sorted. > >But you are halving the number of page cleaners you could configure. In >this example, you'd merely be cutting write performance by 50%. I can't >see the point in doing that. For small writes, I agree. All I'm saying is that the sorted write would still be sorted. >I rest my original case as I've *still* seen nothing here to disprove it:- I wasn't trying to disprove it. In fact I agree with what you say. Basically, LVM has significant advantages for reads, is better than RAID for writes, and is worse than normal chunks for writes. There is a place for RAID and LVM, and you should use all three types where appropriate even on the same instance. As always, it depends on the application. David W Chan Objective Research Ltd dwc@cix.compulink.co.uk