Win-98 on a 500 gb SATA drive

[98 Guy]

I used Western Digital software to create a single FAT-32 partition on a WD5000KS SATA drive - using 4kb cluster size (about 121 million clusters). This is on my Asrock motherboard.

Read here for details:

http://www.msfn.org/board/index.php?showtopic=92792

Anyways, just to let you know, the Windows-ME versions of defrag, scandskw and dskmaint.dll do not work given the rather large number of clusters. DOS scandisk does work though.

By the way, has anyone here ever heard speculation of a win-98 limitation as to the number of files or directories it can have?

[eidenk]

That's what MS says :

The maximum possible number of clusters on a volume using the FAT32 file system is 268,435,445. With a maximum of 32 KB per cluster with space for the file allocation table (FAT), this equates to a maximum disk size of approximately 8 terabytes (TB).

http://support.microsoft.com/kb/184006

[98 Guy]

I've been exploring win-98 limitations with regard to the use of large volumes with smaller-than-intended cluster size, specifically the use of 4kb cluster size on volumes ranging from 24 to 500 gb.

I've posted my results to microsoft.public.win98.gen_discussion - you can read them here:

http://tinyurl.com/33tytp

http://tinyurl.com/2umqoz

http://tinyurl.com/2wf5gk

http://tinyurl.com/2kh8kn

Basically, Microsoft has tried to limit FAT-32 volumes from having more than 2 million clusters, and only when necessary allows the cluster count to rise above 2 million to the documented limite of 4.177 million (this happens for volumes between 64 to 128 gb).

I've found that DOS scandisk will handle an absurdly high cluster count (121 million clusters on a 500 gb volume) but win-98 scandisk and defrag are limited to 4 million clusters. However, windows-ME scandisk and defrag are capable of handling a volume with more than 4 million clusters - but not as many as 121 million.

It seems strange that Microsoft limited win-98 defrag and scandisk to 16-bit functionality - seems like an intentional handicap.

Is anyone here aware of any functional or performance issues for win-98 given the expansion of the size of the FAT table when the cluster size is kept small and the cluster-count grows beyond (or way beyond) the intended maximum of 4.177 million?

[oscardog]

I've been exploring win-98 limitations with regard to the use of large volumes with smaller-than-intended cluster size, specifically the use of 4kb cluster size on volumes ranging from 24 to 500 gb.

I've posted my results to microsoft.public.win98.gen_discussion - you can read them here:

http://tinyurl.com/33tytp

http://tinyurl.com/2umqoz

http://tinyurl.com/2wf5gk

http://tinyurl.com/2kh8kn

Basically, Microsoft has tried to limit FAT-32 volumes from having more than 2 million clusters, and only when necessary allows the cluster count to rise above 2 million to the documented limite of 4.177 million (this happens for volumes between 64 to 128 gb).

I've found that DOS scandisk will handle an absurdly high cluster count (121 million clusters on a 500 gb volume) but win-98 scandisk and defrag are limited to 4 million clusters. However, windows-ME scandisk and defrag are capable of handling a volume with more than 4 million clusters - but not as many as 121 million.

It seems strange that Microsoft limited win-98 defrag and scandisk to 16-bit functionality - seems like an intentional handicap.

Is anyone here aware of any functional or performance issues for win-98 given the expansion of the size of the FAT table when the cluster size is kept small and the cluster-count grows beyond (or way beyond) the intended maximum of 4.177 million?

Excellent research very well done, only problems I am aware of would be decrease in hd performance and higher fragmentation,would be an ideal cluster size for lots of small files though. I do not know if this might be of interest to you but I wonder how ext2 performs with the dos/windows drivers available for download