Re: Don't Fix It if it is Not Broken (was Looking at Macs...)
- From: TheLetterK <theletterk@xxxxxxxxxxxxxxxxx>
- Date: Mon, 22 Aug 2005 21:19:39 -0400
Mark Conrad wrote:
TheLetterK pontificated -
plus the desirability to keep a lot of the app's in the same partition as the OS, makes for a horrendously bigger partition that needs more time consuming maintenance.
OS X performs a LOT more in the way of automated disk maintanence than OS 9 did. Very rarely should you actually need to manually perform disk maintanence.
Agreed, OS X does a lot more in the way of self maintenance.
However...
OS X itself is only a few GBs. If OS X were the only thing in my partition, I would agree with you, that it would rarely need any maintenance other that the self maintenanance it automatically does.
The problem arises when outside app's are added. I added two things, iLife and iWork, plus a few util's like TechTool, so already I am up to 8.5 GBs being used in my 20 GB partition.
Your point?
Only if your using apps that don't call upon OS X to write to the disk. Needless to say, these aren't very common.
Using those outside app's increases the chances for file corruption to occur.
The more app's I use, the greater the chances for corruption,
This simply is not true.
This is not at all the case. OS X (and by this, I include the vast majority of OS X apps) does active disk maintanence to circumvent the issues entirely. I just isn't succeptable to the same kind of disk problems OS 9 was.therefore it makes sense to maintain the OS and the file system more often, because they are both being used more heavily, so therefore more subject to being corrupted.
(whew, that was a long mouthful)
Many things can go wrong on those big partitions. I seem to be spending a lot of my time doing routine maintenance lately.
Then you have something very wrong going on. What, exactly, are you doing and what's causing this to occur?
Naw, nothing "really wrong" going on,
Your getting B-tree errors, this indicates that something is bad wrong.
Just to note, repairing permissions rarely does anything from a practical usage standpoint.merely adding updates to all the various app's, repairing permissions because some of those updates mess up the permissions.
It shouldn't be getting fragmented in the first place. OS X has a wide array of technologies in place to almost completely elimiate the situations where fragmentation is an issue. Unless your continually rewriting large blocks of data at once, you shouldn't be having a problem here.I also take a quick look at the fragmentation of my 20-GB partition, and if it is bad enough, I do a defrag'.
Very much self-inflicted--the majority of the steps you take are unnessesary, or are indicitive of a more serious problem.
I admit a lot of my maintenance work is self inflicted, because of the way I prefer to work here.
pdisk might cause partitioning issues, but nothing there would cause the type of problems you describe (I don't understand how dc might cause a problem of any sort, let alone the disk issues you describe).
Explanation - *********************************** In an effort to learn a little Unix, I find it necessary to "experiment" with using the built in Unix app's that come with OSX, like the commands dd, dc, pdisk, rm, split, chmod, plus all the many "arguments" used to modify how those Unix commands work. Those commands are very dangerous, so often I completely zap my OSX, and the HFS+ file system, and even the partition structures. Stuff gets damaged to the extent where conventional backup utilities will not restore my disk, so I have to use a custom backup scheme.
Why don't you just install the apps your going to use, then backup. Keep user data on a seperate partition or off on a server. When you need to restore just go back to the working system backup.
Okay, so when I add any significant software to my OS X partition, software that needs a lot of tedious configuration and setup in order to work properly, I have to create a new custom backup file.
Your doing a lot of extrenious crap that doesn't seem to be warranted from the usage you describe.
This is a real hassle, involving many hours of "maintaining" OSX, doing stuff like freshly restoring the OSX partition from the "present" backup file, repairing permissions, rebuilding the directory, repairing the file system, defragging the OSX partition, wiping the free space of the OSX partition, creating the new backup itself, then compressing a copy of the new backup file to be stored on a off-site disk.
Your doing a lot more work than any other OS X admin I know--and your usage pattern (from what you've described) doesn't seem to warrant the precautionary work you do.
Then I can settle down to work for a day, week, or month until another significant application demands that I create a newer backup file.
With my newest backup file at-the-ready, I can learn new Unix tricks to my heart's content, secure in the knowledge that I can restore my disk if I mess anything up. ************************************ End of Explanation
(no "journaling" on, by the way, on my 20 GB internal drive)
Why don't you enable journaling? There's no reason to keep it off unless your actively doing disk maintanence.
Agreed, I failed to make that clear, it was a temporary turn-off of journaling. Journaling is valuable because it essentially repairs file system damage shortly after it happens, keeping a bad file system from continually damaging files over a long period of time.
Perhaps a few files incur damage, but not a whole bunch of files.
My main objection to how journaling works in OS X is that it does not _report_ when it "fixes" the file system.
The idea is that these fixes should be transparent.
Your completely misevaluating the situation. IF anything, you should be doing LESS work with OS X than you did OS 9. It's a much, much easier system to maintain. You seem to be doing a hell of a lot of unnessesary work.
If it reported, then at least I would be aware that a file may have been damaged.
are
SURPRISE, massive failure, disk repair gave up, showed a message: "B-tree node invalid value, can't repair"
You might want to check the drive integrity if B-tree errors keep popping up. That is *not* a normal occurance.
I never did find out what caused it. After much diddling, the error disappeared, never to come back again.
Unfortunately, this was all in the wee hours of the morning and I was tired, so I made the cardinal mistake of not recording what I did during my "diddling", so I have no idea why the error cleared up.
I have ran for literally many months on my external OS X drive, without doing any regular maintenance on it. I am a bit more cautious with my main internal drive, adding updates, pestering the drive partitions with rebuilding their directories, low-level disk checking, rebuilding "Volume Structures" and files with TechTool, even an occassional disk defragging episode.
You are aware that you can do more damage with over-caution than with under-caution, yes? It's a case of 'don't fix it if it's not broken'.
Ahh, now at long last we get down to the real meat of the subject line of this post ;-)
It's a case of 'don't fix it if it's not broken'.
Well, yes and no, an alternate saying could be:
"Fix it to keep it from breaking" <g>
I know what you are getting at, though. Excessive testing on its own can easily cause damage.
It is hard to determine what is too little testing, versus what is too much testing.
My rule of thumb is that the more likely something is to fail, the more often I am going to test it.
I seldom test newer disk drives, but I often test older ones.
Same with newer RAM versus old RAM.
RAM has a definite life span, according to IBM (see item "4." in website below -
<http://www.anandtech.com/guides/viewfaq.aspx?i=3>
A file system handling a tremendous amount of file traffic in a very crowded and huge OS X partition is more likely to fail than a lightly used file system, so therefore will need more maintenance attention from utilities like TechTool Pro.
No, no it's not. Mostly because of how well OS X handles disk management.
An old Mac subjected to heat, cold, vibration, humidity, dust, oil fumes, and heavy usage is going to get all sorts of testing from me.
My new 17" powerbook is setting a few feet away from me in its hard case, with fresh silica gel inside. Ya know, I have not ran one test on that in several weeks.<g>
Gotta take it out of its case one of these days, to charge its battery.
Mark-
.
- Follow-Ups:
- Re: Don't Fix It if it is Not Broken (was Looking at Macs...)
- From: Mark Conrad
- Re: Don't Fix It if it is Not Broken (was Looking at Macs...)
- References:
- Don't Fix It if it is Not Broken (was Looking at Macs...)
- From: Mark Conrad
- Don't Fix It if it is Not Broken (was Looking at Macs...)
- Prev by Date: Re: AltiVec goes the way of the DoDo
- Next by Date: Re: Tablet MacDell
- Previous by thread: Re: Don't Fix It if it is Not Broken (was Looking at Macs...)
- Next by thread: Re: Don't Fix It if it is Not Broken (was Looking at Macs...)
- Index(es):
Relevant Pages
|