Proof that Journalling does not defrag

Veljo

Veljo

Mac Enthusiast
I found my Panther boot disk becomming extremely slow, and with journalling enabled there's no way to defrag. So, I opened up Disk Utility and disabled journalling (power outages are extremely rare and when they do happen I'll take my chances, I'd rather have a faster computer) and then proceeded to defrag using Drive 10. I was shocked to see the condition of my hard disk: not only was it fragged bad but I had to repair the volume structures before it could even do so. Check out the pic attached.
 

Attachments

  • notdefragged.jpg
    notdefragged.jpg
    86.9 KB · Views: 70
Veljo said:
I found my Panther boot disk becomming extremely slow, [snip] I'd rather have a faster computer) and then proceeded to defrag using Drive 10.[snip].
Click to expand...

Veljo, I was under the impression that Drive 10 did not work under Panther yet? I was waiting for an update.

How did it work for you?
 
For what its worth, I think that looks more like file system fragmentation. AFAIK, Panther's auto defragmentation works on a file by file basis, so it won't defrag the entire filesystem (i.e. place commonly used files together, etc), it will ensure that a file that is under 20 MB doesn't get placed on sectors that are too far apart.
 
Journaling does not do any type of de-frag, just keeps track of writes to the HD (the journal!). Leaving journaling enabled, as viro stated, also allows the 'Adaptive Hot-file Clustering' that Darwin provides (which is not a full-drive de-frag, just an optimization of certain file types and sizes). I have also read that, despite the slightly slower writes, system performance with journaling enabled actually is measurably (very little) faster on some operations, or no real difference.
228 file fragments out of 80-some thousand total files is not what I would call massive framentation, also a 4 GB partition for OS X is really tight, and free space would be critical, especially if you have around 1 GB or less free space. Is this on your 60 GB iMac? How did you end up with such a small partiton for OS X?
 
As said before, journalling and defragging are not the same thing, and are not related at all. Mac OS X 10.3 has a journalled file system, as well as defragmentation routines. I was under the impression that both worked simultaneously, i.e., adaptive hot-file clustering was still in effect even with journalling enabled.

Veljo: I don't mean to sound rude or anything, but I think that you're relying too heavily on defragmenting your OS X system. I think you'll find quickly that your drive gets fragmented again soon after you defrag, simply because of the 1,000s of files that OS X uses on a daily basis. Defragmenting your drive under OS X is not nearly as beneficial to the system as it was under OS 9.

Just for kicks, how much of a speed boost are you seeing after you defragment? Does it really provide any noticeable increase in speed, or does it just SEEM like it's faster?

i used to defragment my drive in OS X 10.0.x and 10.1.x, but the speed boost I thought I saw was just not there. It really didn't help other than giving me peace of mind that I had done SOMEthing to help the system, which I really wasn't doing at all. I think you'll find that if you run the optimization program again in two days that your drive has returned to the fragmented state you found it in when you ran the optimization. OS X puts files on the disk where it likes to, and defragmenting just moves those files away from there, where OS X is just gonna put them back anyway.
 
fragmentation is a farce, it's all a conspiracy....Heh :p I use discwarroir every now and again to defragment and i'm not sure it actually speeds up anything...but eh...it's like taking a sugar pill ;) Placebo central baby!
 
Under 10.3, the OS defrags only files under 20 MB, only if they are spread over more than some minimum number of fragments (maybe 8? I forget), and only when you access the file. It's not meant to be a full replacement for a defrag utility, just a nice little performance tweak, where we'll see the benefits gradually show up over time, as our frequently used files become somewhat less fragmented.

Files that have only a small number of fragments, or that are over 20 MB, or that you have not read or written since turning on journalling, will not have been defragmented. So your results are not surprising at all - the OS basically only defrags files where the fragmentation would make a major difference to performance, not by any means all of them.
 
Partition size seems like a more likely suspect for the lack of speed. Even if you only have the basic system on that partition (not likely since you are using 3.2 gig) it seems pretty small. 30% free space...
 
I've read all of your opinions and now I agree. So, what do you all suggest? My hard disk is as said 4GB and 56GB partitions...what should I allocate to the OS X partition? 10GB?
 
I would go even larger. With such a big drive, I would use 30GB, unless what you have on the other partition cannot be shared with the OSX volume. This will give a significant boost in performance.
 
IMHO, with 'just' a 60 GB hard drive, I would split the difference, if you're going to partition, use around 15 GB for the boot partition, you're already using your user folders on a different partition, and that's where your largest files will be, and 15 GB (more or less) will give you plenty of room if you need to do an 'archive and install' with your system. There's not enough space to do that option on your present 4 GB boot partition.
 
DeltaMac said:
also a 4 GB partition for OS X is really tight, and free space would be critical, especially if you have around 1 GB or less free space. Is this on your 60 GB iMac? How did you end up with such a small partiton for OS X?
Click to expand...

I was going to ask the same question, what is up with that?
 
I use a 60GB as my main drive, had it split equally and wiped it for Panther without any partitions and it runs very well.
 
You must log in or register to reply here.
Back
Top