Maggot061998
asked on
Cluster sizes wrong on NTFS HDD in new WinXP-h. What should I do?
I built my own Pentium4-1.8A / RDRAM machine. I formatted and partitioned the Seagate 40 GB 7200 rpm into C: ("OS")
10 GB and D: ("Data") 30 GB. I'd planned the format to be the NTFS system with 8 kB clusters on the 10 GB drive and 16 kB clusters for the 30 GB partition.
After polling the experts here last week, I decided to reorganize my usage of my partitions so that the D: received apps. We briefly considered wisdom of creating a third partition for Temp in order to keep the most fragmented directories sequestered.
There was a consensus that I should go ahead and buy Partition Magic 7 . OK, I got it. One of the first things I did was to look at its info on my drives and I was shocked to discover that the cluster sizes were not what I'd thought I'd made: On the C: (10GB) clusters are 4kB and on D: (30GB) clusters are 512 B. (I reprint the analysis from PM below).
The operation of Resizing Clusters in the Advanced Operation menu of PM appears unavailable (grayed out) to both my partitions as best as I can tell.
What are the consequences of leaving these clusters sized as they are? Do you advise me to resize them? How?
Does one lose data in cluster resizing?
THanks,
DEan
=====
PowerQuest PartitionInfo 7.0 -- Windows NT/2000 Version
Date Generated: 03/24/02 22:41:10
Copyright (c)1994-2001, PowerQuest Corporation
Permission is granted for this utility to be freely copied so long
as it is not modified in any way. All other rights are reserved.
PowerQuest, makers of PartitionMagic(r), Drive Image(tm), and DriveCopy(tm), can be reached at:
Voice: 801-437-8900
Fax: 801-226-8941
Web site: http://support.powerquest.com/support.html
E-mail: magic@powerquest.com
General System Information:
Total Physical Memory (bytes): 536,309,760
Used Physical Memory: (bytes): 147,283,968
Maximum Page File Size: (bytes): 1,311,436,800
Current Page File Size: (bytes): 98,095,104
==========================
Disk Geometry Information for Disk 1: 4865 Cylinders, 255 Heads, 63 Sectors/Track
System PartSect # Boot BCyl Head Sect FS ECyl Head Sect StartSect NumSects
==========================
0 0 80 0 1 1 07 1023 254 63 63 19,534,977
Info: End C,H,S values were large drive placeholders.
Actual values are:
0 0 80 0 1 1 07 1215 254 63 63 19534977
0 1 00 1023 0 1 0F 1023 254 63 19,535,040 58,621,185
Info: Begin C,H,S values were large drive placeholders.
Info: End C,H,S values were large drive placeholders.
Actual values are:
0 1 00 1216 0 1 0F 4864 254 63 19535040 58621185
19,535,040 0 00 1023 1 1 07 1023 254 63 19,535,103 58,621,122
Info: Begin C,H,S values were large drive placeholders.
Info: End C,H,S values were large drive placeholders.
Actual values are:
19535040 0 00 1216 1 1 07 4864 254 63 19535103 58621122
==========================
Partition Information for Disk 1: 38,162.2 Megabytes
Volume PartType Status Size MB PartSect # StartSect TotalSects UsedSects FreeSects
==========================
C: NTFS Pri,Boot 9,538.6 0 0 63 19,534,977 19,534,977 0
ExtendedX Pri 28,623.6 0 1 19,535,040 58,621,185 58,621,185 0
EPBR Log 28,623.6 None -- 19,535,040 58,621,185 58,621,185 0
D: NTFS Log 28,623.6 19,535,040 0 19,535,103 58,621,122 58,621,122 0
==========================
Boot Record for drive C: (Drive: 1, Starting sector: 63, Type: NTFS)
==========================
1. Jump: EB 52 90
2. OEM Name: NTFS
3. Bytes per Sector: 512
4. Sectors per Cluster: 8
5. Reserved Sectors: 0
6. Number of FATs: 0
7. Root Dir Entries: 0
8. Total Sectors: 0
9. Media Descriptor: 0xF8
10. Sectors per FAT: 0
11. Sectors per Track: 63 (0x3F)
12. Number of Heads: 255 (0xFF)
13. Hidden Sectors: 63 (0x3F)
14. Total Sectors (>32MB): 0 (0x0)
15. Unused: 0x80008000
16. Total NTFS Sectors: 19534976
17. MFT Start Cluster: 786432
18. MFT Mirror Start Clust: 1220936
19. Clusters per FRS: 246
20. Clusters per Index Blk: 1
21. Serial Number: 0x9E9C54F79C54CB81
22. Checksum: 0 (0x0)
23. Boot Signature: 0xAA55
==========================
Boot Record for drive D: (Drive: 1, Starting sector: 19,535,103, Type: NTFS)
==========================
1. Jump: EB 52 90
2. OEM Name: NTFS
3. Bytes per Sector: 512
4. Sectors per Cluster: 1
5. Reserved Sectors: 0
6. Number of FATs: 0
7. Root Dir Entries: 0
8. Total Sectors: 0
9. Media Descriptor: 0xF8
10. Sectors per FAT: 0
11. Sectors per Track: 63 (0x3F)
12. Number of Heads: 255 (0xFF)
13. Hidden Sectors: 19535103 (0x12A14FF)
14. Total Sectors (>32MB): 0 (0x0)
15. Unused: 0x80008000
16. Total NTFS Sectors: 58621121
17. MFT Start Cluster: 6291456
18. MFT Mirror Start Clust: 29310560
19. Clusters per FRS: 2
20. Clusters per Index Blk: 8
21. Serial Number: 0x881C590B1C58F5A2
22. Checksum: 0 (0x0)
23. Boot Signature: 0xAA55
10 GB and D: ("Data") 30 GB. I'd planned the format to be the NTFS system with 8 kB clusters on the 10 GB drive and 16 kB clusters for the 30 GB partition.
After polling the experts here last week, I decided to reorganize my usage of my partitions so that the D: received apps. We briefly considered wisdom of creating a third partition for Temp in order to keep the most fragmented directories sequestered.
There was a consensus that I should go ahead and buy Partition Magic 7 . OK, I got it. One of the first things I did was to look at its info on my drives and I was shocked to discover that the cluster sizes were not what I'd thought I'd made: On the C: (10GB) clusters are 4kB and on D: (30GB) clusters are 512 B. (I reprint the analysis from PM below).
The operation of Resizing Clusters in the Advanced Operation menu of PM appears unavailable (grayed out) to both my partitions as best as I can tell.
What are the consequences of leaving these clusters sized as they are? Do you advise me to resize them? How?
Does one lose data in cluster resizing?
THanks,
DEan
=====
PowerQuest PartitionInfo 7.0 -- Windows NT/2000 Version
Date Generated: 03/24/02 22:41:10
Copyright (c)1994-2001, PowerQuest Corporation
Permission is granted for this utility to be freely copied so long
as it is not modified in any way. All other rights are reserved.
PowerQuest, makers of PartitionMagic(r), Drive Image(tm), and DriveCopy(tm), can be reached at:
Voice: 801-437-8900
Fax: 801-226-8941
Web site: http://support.powerquest.com/support.html
E-mail: magic@powerquest.com
General System Information:
Total Physical Memory (bytes): 536,309,760
Used Physical Memory: (bytes): 147,283,968
Maximum Page File Size: (bytes): 1,311,436,800
Current Page File Size: (bytes): 98,095,104
==========================
Disk Geometry Information for Disk 1: 4865 Cylinders, 255 Heads, 63 Sectors/Track
System PartSect # Boot BCyl Head Sect FS ECyl Head Sect StartSect NumSects
==========================
0 0 80 0 1 1 07 1023 254 63 63 19,534,977
Info: End C,H,S values were large drive placeholders.
Actual values are:
0 0 80 0 1 1 07 1215 254 63 63 19534977
0 1 00 1023 0 1 0F 1023 254 63 19,535,040 58,621,185
Info: Begin C,H,S values were large drive placeholders.
Info: End C,H,S values were large drive placeholders.
Actual values are:
0 1 00 1216 0 1 0F 4864 254 63 19535040 58621185
19,535,040 0 00 1023 1 1 07 1023 254 63 19,535,103 58,621,122
Info: Begin C,H,S values were large drive placeholders.
Info: End C,H,S values were large drive placeholders.
Actual values are:
19535040 0 00 1216 1 1 07 4864 254 63 19535103 58621122
==========================
Partition Information for Disk 1: 38,162.2 Megabytes
Volume PartType Status Size MB PartSect # StartSect TotalSects UsedSects FreeSects
==========================
C: NTFS Pri,Boot 9,538.6 0 0 63 19,534,977 19,534,977 0
ExtendedX Pri 28,623.6 0 1 19,535,040 58,621,185 58,621,185 0
EPBR Log 28,623.6 None -- 19,535,040 58,621,185 58,621,185 0
D: NTFS Log 28,623.6 19,535,040 0 19,535,103 58,621,122 58,621,122 0
==========================
Boot Record for drive C: (Drive: 1, Starting sector: 63, Type: NTFS)
==========================
1. Jump: EB 52 90
2. OEM Name: NTFS
3. Bytes per Sector: 512
4. Sectors per Cluster: 8
5. Reserved Sectors: 0
6. Number of FATs: 0
7. Root Dir Entries: 0
8. Total Sectors: 0
9. Media Descriptor: 0xF8
10. Sectors per FAT: 0
11. Sectors per Track: 63 (0x3F)
12. Number of Heads: 255 (0xFF)
13. Hidden Sectors: 63 (0x3F)
14. Total Sectors (>32MB): 0 (0x0)
15. Unused: 0x80008000
16. Total NTFS Sectors: 19534976
17. MFT Start Cluster: 786432
18. MFT Mirror Start Clust: 1220936
19. Clusters per FRS: 246
20. Clusters per Index Blk: 1
21. Serial Number: 0x9E9C54F79C54CB81
22. Checksum: 0 (0x0)
23. Boot Signature: 0xAA55
==========================
Boot Record for drive D: (Drive: 1, Starting sector: 19,535,103, Type: NTFS)
==========================
1. Jump: EB 52 90
2. OEM Name: NTFS
3. Bytes per Sector: 512
4. Sectors per Cluster: 1
5. Reserved Sectors: 0
6. Number of FATs: 0
7. Root Dir Entries: 0
8. Total Sectors: 0
9. Media Descriptor: 0xF8
10. Sectors per FAT: 0
11. Sectors per Track: 63 (0x3F)
12. Number of Heads: 255 (0xFF)
13. Hidden Sectors: 19535103 (0x12A14FF)
14. Total Sectors (>32MB): 0 (0x0)
15. Unused: 0x80008000
16. Total NTFS Sectors: 58621121
17. MFT Start Cluster: 6291456
18. MFT Mirror Start Clust: 29310560
19. Clusters per FRS: 2
20. Clusters per Index Blk: 8
21. Serial Number: 0x881C590B1C58F5A2
22. Checksum: 0 (0x0)
23. Boot Signature: 0xAA55
Unlike a FAT16 (or, to some extent, a FAT32 partition) the cluster size on an NTFS drive does not need to get bigger as the drive size does--you could have a terabyte NTFS partition with 512 byte clusters if you so chose. The only consequence of having small clusters is that drive fragmentation tends to get worse because more files are split into multiple clusters.
ASKER
"The only consequence of having small clusters is that drive fragmentation tends to get worse because more files are split into multiple clusters."
My largest (30 GB) drive is at 512B/cluster instead of optimal 16 kB/cluster.
The cluster size, at roughly 3% of ideal, is going to fragment more than otherwise. I understand this much, PJ.
I need to get a feel for the pros & cons of correcting my error. Please give me a feel for my options:
1) Assuming I were determined to change cluster size on my 30 GB partition, how
would I go about it (broad strokes) and what would be consequences in the short term- data loss?
2) About how much inefficiency or performance hit do I incur if I leave the cluster sizes mismatched as they currently are ?
Thanks (& good night!)
My largest (30 GB) drive is at 512B/cluster instead of optimal 16 kB/cluster.
The cluster size, at roughly 3% of ideal, is going to fragment more than otherwise. I understand this much, PJ.
I need to get a feel for the pros & cons of correcting my error. Please give me a feel for my options:
1) Assuming I were determined to change cluster size on my 30 GB partition, how
would I go about it (broad strokes) and what would be consequences in the short term- data loss?
2) About how much inefficiency or performance hit do I incur if I leave the cluster sizes mismatched as they currently are ?
Thanks (& good night!)
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
You can try PM 7 (partition magic 7) from www.powerquest.com
ASKER
Steven wrote:
" You can try PM 7 (partition magic 7) from www.powerquest.com"
Dean had posted originallY:
"There was a consensus that I should go ahead and buy Partition Magic 7 . OK, I got it. One of the first things I did was to look at its info on my drives and I was shocked to discover that the cluster sizes were not what I'd thought I'd made: On the C: (10GB) clusters are 4kB and on D: (30GB) clusters are
512 B. (I reprint the analysis from PM below).
The operation of Resizing Clusters in the Advanced Operation menu of PM appears unavailable (grayed out) to both my partitions as best as I can tell."
Dean to Steve:
Hey you!
You know I value your advice (<Gasp!>) so, sarcasm aside (<GASP!!>), do you concur that -even with PM7- the only way to resize cluster will involve a reformat with data loss?
Do you also agree that -even with a pretty new pc whose could be reconstituted
in about 6 hours (my guess), that there is little or no performance hit I'd incur if I left my NTFS HDD with such small cluster size as these (4kB for 10 GB partition and 512 bytes for the 30 GB partition) ??
" You can try PM 7 (partition magic 7) from www.powerquest.com"
Dean had posted originallY:
"There was a consensus that I should go ahead and buy Partition Magic 7 . OK, I got it. One of the first things I did was to look at its info on my drives and I was shocked to discover that the cluster sizes were not what I'd thought I'd made: On the C: (10GB) clusters are 4kB and on D: (30GB) clusters are
512 B. (I reprint the analysis from PM below).
The operation of Resizing Clusters in the Advanced Operation menu of PM appears unavailable (grayed out) to both my partitions as best as I can tell."
Dean to Steve:
Hey you!
You know I value your advice (<Gasp!>) so, sarcasm aside (<GASP!!>), do you concur that -even with PM7- the only way to resize cluster will involve a reformat with data loss?
Do you also agree that -even with a pretty new pc whose could be reconstituted
in about 6 hours (my guess), that there is little or no performance hit I'd incur if I left my NTFS HDD with such small cluster size as these (4kB for 10 GB partition and 512 bytes for the 30 GB partition) ??
ASKER
Friends,
A pal who scanned the PM data I posted above asked me: Why CHS?
Why indeed. I see that the BIOS has Translation Method of "Match Partition Table". This really needs to be changed to to LBA. This may pre-empt my cluster size concern- if I need to reformat after changing to LBA anyway.
Clearly, I suck at HDD formatting. Is it down to that now: that I just need to start over from scratch?
Dean
A pal who scanned the PM data I posted above asked me: Why CHS?
Why indeed. I see that the BIOS has Translation Method of "Match Partition Table". This really needs to be changed to to LBA. This may pre-empt my cluster size concern- if I need to reformat after changing to LBA anyway.
Clearly, I suck at HDD formatting. Is it down to that now: that I just need to start over from scratch?
Dean
Well Dean, when youchange the bios to LBA, hopefully it will boot right up and youwon't have to worry about it. I suggest using the autodetect feature of the bios. It usually sets it cvorrectly
ASKER
Thanks, PJ and Steven, as this seems to be one of those situations where no one can predict the optimal course or results I'll just back-up, change the BIOS translation method to LBA, and see what happens.
If it boots okay- I'll still have bittersweet feeling about the screwed up cluster size: esp. the 512 bytes all through the 30 GB partition. It seems counter-intuitive to me that
forcing the HDD to read so many cluster each time it's seeking would not impact its performance. Since the machine is so young, I'd regret hamstring it from its birth.
One pal has hinted an easier path than starting from scratch. It involved using a utility I don't yet have ("Ghost") to copy my data onto a partition (that I might create with my PM7) which -if I understand correctly and probably do not- would then free up the empty partition so I could change cluster size there.
Any reaction to the advisability or plausibility of sparing me lots of reistallation work using Ghost would be appreciated.
Thanks, guys.
Dean
If it boots okay- I'll still have bittersweet feeling about the screwed up cluster size: esp. the 512 bytes all through the 30 GB partition. It seems counter-intuitive to me that
forcing the HDD to read so many cluster each time it's seeking would not impact its performance. Since the machine is so young, I'd regret hamstring it from its birth.
One pal has hinted an easier path than starting from scratch. It involved using a utility I don't yet have ("Ghost") to copy my data onto a partition (that I might create with my PM7) which -if I understand correctly and probably do not- would then free up the empty partition so I could change cluster size there.
Any reaction to the advisability or plausibility of sparing me lots of reistallation work using Ghost would be appreciated.
Thanks, guys.
Dean
Ghost will certainly archive an entire partition and then restore it--the only thing I don't know is if it'll be happy that the cluster size has changed on the partition between times. I suspect Ghost (when backing up a partition rather than an entire drive) works at a file level, so it ought to be OK, but I have no evidence for this view other than a gut feeling.
http://www.pcguide.com/ref/hdd/file/ntfs/archCluster-c.html
http://www.pcguide.com/ref/hdd/file/ntfs/arch-i.htm
http://www.win2000mag.com/Articles/Index.cfm?IssueID=27&ArticleID=3455
http://searchwin2000.techtarget.com/sDefinition/0,,sid1_gci212675,00.html
And this one is very interesting:
http://www.wec.ufl.edu/staff/hydew/comp/os/nt_ntfs.html