Amiga.org
Amiga computer related discussion => General chat about Amiga topics => Topic started by: HellCoder on September 20, 2006, 05:10:53 PM
-
Hi,
I've always used Reorg on my A1200 when I wanted to reorganise my HD and I almost used it on my current setup. I suddenly realised, does it actually work on my OS3.9 partition with FFS installed on it when the partition is crossing the 4 GB border and larger dan 4 Gb ???
Anyone who can confirm this ?
-
does it actually work on my OS3.9 partition with FFS installed on it when the partition is crossing the 4 GB border and larger dan 4 Gb ?
No.
-
DiskMonTools.lha is the only one that works - but it's slow!
-
in my setup it's not working even this.
give sfs (http://strohmayer.org/sfs/files/SFS_1.254_68k.lha) a try. it has a cool and fast defrag tool too.
-
I would stay well clear of ReOrg... even when patched to v3.11.
The damn thing used to corrupt my disks, crash etc.
It's not worth risking all your data to speed it up a tiny fraction.
It would interest me if, in real world usage, defragmentation actually speeds up reading and writing. Putting all the data in one chunk at the start of a disk as opposed to it being spread over the drive might even create wear on one particular area of a disk over time.
What do others think?
-
that's a new idea for me and not bad at all... in amiga it's not important the defrag i agree, but in video systems and a lot of gb avi files defrag is a must otherwise the droped frames experiance during videotape recording is a pain...
-
@Hyperspeed
The drive doesn't actually mind re-reading and writing the same old spot over and over again. The platter stack spins anyway and the heads don't (really) have to move.
However, it does mind the read/write head being moved like crazy all over the platter when fragmentation is high - just listen to the sound the drive makes: this is certainly not exactly conserving its life.
Depending on the usage of the drive, overdoing on defragmentation might put more stress on the drive than living with fragmentation as you've pointed out. If the drive has a relatively high stress level (e.g. Windows system drive), it's good to have the static files defragmented to reduce the stress on the mechanics. If your drive works as an archive for text documents and such, there's no point in defragmenting it over and over again. Usually it's a good idea to separate the system files and temporary data on different partitions, so the latter cannot fragment the former. (Of course, only putting new files can get them fragmented, so if you don't change any of the files regularly needed, there's nothing to be feared.)
-
by Zac67:
The drive doesn't actually mind re-reading and writing the same old spot over and over again.
Well a drive's life isn't measured in years or how much it's head has wobbled about... it's in write cycles right?
I once created a bad block on a Win 3.11 system when I yanked the plug out just as it was reading a file. Possibly it was also writing to a swap file or something but it goes to show that the disk surface can be affected adversely by the head.
Bad blocks are usually mapped out at the manufacturer but they do appear, who knows why. Maybe it's to do with the magnetic coating losing it's gauss or whatever they measure it in.
The one thing that I have seen a drive or two dying from though is dead bearings. I have a sneaking suspicion that mounting a drive sideways or upwards may cause bearing wear... but I have only seen it on Windows machines. Maybe the constant writing to virtual memory and endless bloated files on bootup just make the drive commit suicide!
:-D
By all means use ReOrg to see how fragmented a drive is but I would advise using a UPS (uninteruptable power supply), a good IDE/SCSI cable and remove all hacks/tasks before use.
DiskSalv will not help you if ReOrg has a tantrum... you will lose everything.
-
DiskSalv will not help you if ReOrg has a tantrum... you will lose everything.
Never lost anything using ReOrg hundreds of times (good luck with wall power...). And it was very picky to throw out of the process if even a minor, otherwise invisible fault did appear in the system .
-
@Hyperspeed
Well a drive's life isn't measured in years or how much it's head has wobbled about... it's in write cycles right?
Zac67 is right here, excessive seeking is much worse than writing the same spot over and over again. HDDs have practically unlimited "write cycles" (it's electromagnetic after all, nothing physical involved), but read/write head (as mechanical component) has limited life-span.
I once created a bad block on a Win 3.11 system when I yanked the plug out just as it was reading a file. Possibly it was also writing to a swap file or something but it goes to show that the disk surface can be affected adversely by the head.
All new HHDs park the head gracefully if the power is lost. No bad blocks or head crashing will occur. This has been the standard feature for at least 15 years (just the implementation has changed).
Bad blocks are usually mapped out at the manufacturer but they do appear, who knows why.
If you see any bad block with modern HDD, then the HDD is dead already. All modern HDDs have upto 20% "extra" spare storage to replace bad blocks/tracks with. The remapping is fully automagic. Only after this all fails (running out of spare areas etc, or some catastrophic failure occurs) the drive will report the error to host.
All modern drives support S.M.A.R.T (http://en.wikipedia.org/wiki/Self-Monitoring%2C_Analysis%2C_and_Reporting_Technology) for reporting the failures as they happen. With SMART devices and SMART capable monitor application (such as HDD Health (http://www.panterasoft.com/) for Windoze, or smartmontools (http://smartmontools.sourceforge.net/), multiplatform) it is possible to monitor the device health status. This way you will sometimes be able to predict he HDD failure. Sometimes this doesn't help either, the HDD just dies instantly.
-
I'm surprised nobody likes ReOrg because I've used it alot and has never failed me. In my A600 I have a HD that makes alot of hacking noise when swapping tracks. After a reorg loading files won't make that hacky noise coz all tracks are after one another. Hence I used it alot.
But with the multitasking environment you're loading more files at once and you still hear the head swapping tracks. But I'm surprised that my computer at work always gets messed up even if I never delete files. You would expect the files there to stay in place but they get fragmented too ?
Anyway... I've formatted another partition and copyd all files to that. This works too! :-D
-
HellCoder wrote:
I'm surprised nobody likes ReOrg because I've used it alot and has never failed me.
nobody? back in the days of workbench 3.1 i used to use a lot the qbtools defrag tool but when i patched the ffs to 64bit the tool stopped working. after a little search i discovered reorg and used it a lot too. i don't remember if it was working in workbench 3.5 but it's sure that never worked from the day i install workbench 3.9. after a lot of troubles for years with the latest ffs i moved to sfs and for a year now i feel grate :-)
-
It might be faster to defrag by copying all files to other HD, format original and copy all files back. As a by-product you get full backup too :)
And BTW. Isn't the limit for partitions with old defrag programs 2GB. As old FFS limits are 2GB/partition and 4GB/HD. And they probably work only with 512 byte blocksizes.
-
by pVC:
It might be faster to defrag by copying all files to other HD, format original and copy all files back. As a by-product you get full backup too :)
I tried that once but ReOrg reported the destination drive was just a clone of the fragmented source. The fragmentation had copied too. This may be FFS though.
Piru: Do you know anything on magnetic hysteresis, I'm curious as to how this works.
Also, modern drive reporting tools are not available for Amiga and HDToolBox can't even perform a CheckDisk as efficiently as the PC.
AmigaOS 4.0's new HDToolBox looks neat, especially with it's SCSI specific verification features. Wonder if there's a dedicated defragger...
-
@Hyperspeed: hysteresis is a greek word and in general means ''delay'' if this can help you someway. in extream situations can mean ''loss'' too.
-
Thanks amije that's interesting. Magnetic Hysteresis must then mean a delay or loss in magnetism.
From what I've read it's a way of recovering data off a hard disk's surface even after it has been full-formatted multiple times.
Apparently spy agencies use this method but it may be increasingly difficult now with higher densities of data on the drive.
How the process works is a complete mystery to me though...
-
Quite simple - the magnetical particles want to keep their orientation (hysteresis describes that willingness). If you turn the magnetic orientation, slight remnants stay from the preceeding state and it can be read with precision devices. So you have to turn your 0s to 1s 5-10-20 times randomly until you can be quite sure nobody can reproduce the original state - speaking of hard disk surfaces, simply taken.
-
Do you know anything on magnetic hysteresis, I'm curious as to how this works.
I don't, but my friend wikipedia (http://en.wikipedia.org/wiki/Hysteresis#Magnetic_hysteresis) does.
Also, see Data recovery (http://en.wikipedia.org/wiki/Data_recovery) wiki.
A common misconception is that overwritten data ... can be recovered by different means, especially by intelligence agencies.
-
A common misconception is that overwritten data ... can be recovered by different means, especially by intelligence agencies.
A common misconception is that Wikipedia is always right and that intelligience agenices never hide what they know - nor do they put false information into the public domain to reinforce the police state which they strive to attain.
:-D