Wrong size for a hard disk in Win2k

Gentle people,

Again having trouble with my hd's ...

There are two 200GB SATA disks in my machine. BIOS reports for both the same size - 200GB. Win2k as well as PartitionMagic 8.0 report for one 190779MB, for the other 131069MB. Both of them used to have the same - 190770MB - size in Win2k until yesterday when I was ripping happily all my audio cds onto this drive until - blue screen - kernel stack overflow - and my machine was gone. At this time the drive had 3 partitions: Win-Audio 162GB (the ripped audio files) Win-Swap 4GB (only the Windows swap dataset is there) and Win-Data 20GB (all my private data). After restart Win-Audio and Win-Data partition were still there and I could rescue all data from there to the other 200GB drive, but Win-Swap partition was gone for good and Win-Data partition showed a size of 20000 GB! No problem I thought, delete all this **** and start anew and went to PartitionMagic ... When entering this program I got the message "Partition table on drive 5 contains errors, do you want me to fix it?" Aha, I thought may be there is a better way and said yes. Ever since the drive has only 2 thirds of its original capacity ...

The obvious question is - how do I get back the original size? However there are more question when I look at this problem ... How can a Windows swap dataset grow out of any proportion without reporting it to the user? (the swap dataset on my machine is allowed a size from 1534MB to 3000MB) and even if it grows out of any proportion - how can it destroy the partition table of a drive and the partition description of a completely innocent partition?? Or am I wrong and something else is going on which I don't understand????

Any help for the obvious question will be welcome, explanations to the others very much appreciated!!

e.g.gru

 

I'm not really sure exactly what wacked your partition table but it is basically like any other part of the drive except that I think there are two copies kept in completely different parts of the drive so damage or a glitch to a portion of the drive surface should not have messed up like that. However, at this point the damage seems to be done and rebuilding seems like a great idea.

A suggestion to you though. Assuming the "Windows swap dataset" is your pagefile, make the partition a little over twice the max size of the pagefile and use http://www.sysinternals.com/Utilities/PageDefrag.html (free and excellent) to keep it in good condition.

 

Thanks a lot for the explanations! Your assumption is correct - Win-Swap contains just the pagefile. (Being a retired OS/390 system support man it still is the system swap drive to me. Sorry about that ...)

My system is rebuilt and up and running with the knackered hd completely empty. I don't want to integrate it back into my system with only 2/3s of its original capacity. BIOS shows the correct capacity, Windows does not! Is there a tool to completely remove the partition table? Or will a LLF help? And if - does a tool exist to do the LLF??

E.G.Gru

 

Newt,

I understand that defragging a swap file is not possible by ordinary means,
but e.g. gru's setup makes the point moot. His dedicated swap partition is huge.

To keep it from ever fragmenting again in the future, all he has to do is recreate the swapfile with fixed size from the very start:

1) get into system properties and move the swap space to any other drive and reboot

2) go to his swap partition and delete the old swap file (now unused, not locked).

3) get into system properties again and move the swap space to the dedicated drive, assigning it a fixed, and large, size, and reboot.

[ 4) He may have to disable the full-disk warning if pagefile size is set close to partition size. It can be done in registry, but I use Xteq X-setup. ]


This way Windows will never again grow/shrink the swap space, the file will be created contiguous, and stay so forever.

Variable paging file size is a compromise! Size alteration makes sense only if you have modest space available and all data and apps and OS share the same partition.

A nailed-down pagefile relieves the OS from engaging in unneeded gyrations. One of my fav sources of fun is seeing the giant promo screens at my grocery store all frozen reporting that Windows XP is "adjusting the virtual memory file size" or some such nonsense.

No OS/390 here (deep bow) but a bit of Linux, which also teaches some tricks!

Now, I also heard the theory that a pagefile split up and interleaved with all the rest of the HD may make better performance sense due to the VM logic that optimizes HD head travel... but my experience tells me to be very wary of hidden, inscrutable, unchangeable, and undocumented optimizations coming out of the gut of Microsoft.

 

Gentle people,

Just to let you know the end of the story ...
Found "Diskwizard" from Seagate support, installed and executed it and the drive is back to its normal size!

When I come to think of it I still have a hard time to imagine what had happened. The machine shuts down practically every day when I go to bed and is restarted in the morning ---> a new pagefile every day, i.e. fragmentation shouldn't be really a problem. Usually my personal data partition (very low activity) and my swap partition (high activity) reside on a separate drive to have an almost dedicated read/write head for paging. Unfortunately I had to move these partitions onto the large disk because the original page drive started to develop electronic problems (was seen / not seen in BIOS, vanished from the running machine, etc.). Now the page drive has been replaced and the two partitions are back again on their own drive.

Considering everything I saw when the error occurred I can only believe the pagefile as the source of my problem if the program used for ripping did not free up all memory acquired and this in turn led to an explosive growth of the pagefile with all the troubles afterwards. (Still - it shouldn't damage the partition, nor the partition table ...???)

Still puzzled ...

E.G.Gru

 

Agreed - mostly. But I still like pagefiledefrag with NT versions up through 2K (XP doesn't need it) because it deals with some registry hives that are not a part of the pagefile/swapfile. Puts them into a single piece again and crunches out some empty spaces.

I will note that I have tested a printer server that also had a dedicated partition for the swap file and tried both creating new and running the pagedefrag utility and liked the performance better after using it. Also nice on a server machine where you like uptime to be able to get the utility to do an analysis to see if you really need a reboot, regardless of how you deal with the swap file.

 

Newt:

> it deals with some registry hives

Another Big Can of Worms! I saw my registry grow from 65,000 to 90,000 values in two years... I do some cleaning with regcleaner but never defragged it! I should, I will... thank you!