Time Machine Needs too much space, goes too slow?

Barry Lyga wrote:

. . .

Also saw a slew of zero-byte .plist.lockfiles, but my googling around indicates that they're OK to leave.

Yeah, those are left over from Lion, no longer used by Mountain Lion. You can keep them or delete them.

Just rebooted into Safe Mode, then back into normal mode. So I'll give it a day to clear away anything funky and see what's what.

If it detected a problem with the File System Event Store, the Safe Boot would have cleaned it out and the results would be seen immediately (the first backup might take a while, but thereafter should be normal). We've seen that on some Snow Leopard installations, so I doubt that's what's going on here, but the symptoms are similar, and you never know.

Barry Lyga wrote:

. . .

Then, this morning, things slowed down again, and the discrepancy once again became huge. This backup took close to 20 minutes, with half of that being another looooong clean-up routine.

The long clean-up may be temporary -- when the backups made before the Safe Boot are gone -- if the new ones aren't infected with whatever the problem was, their deletions may be faster.

But your backups are still not back to a reasonable speed.

I feel like we've run out of tricks, short of wiping one or both drives.

Mostly. 😟 The only one left is one that did work for a few folks on Snow Leopard -- deleting the File System Event Store completely.

I can't put my hands on a second drive for a while, but once I do, I'll give that a shot and see what's what. (I admit, though, that it seems the problem is that number of files, and that swapping out drives wouldn't make a difference, in that case.)

It does seem unlikely, but odd things do happen. A slowing drive can cause looong backups, of course, but doubtful to cause the high number of files.

Just to confirm, you did delete the Spotlight index earlier, right? If not, that's worth a try (the pink box in #D2 ofTime Machine - Troubleshooting).

Barry Lyga wrote:

Yup, deleted the Spotlight index on the backup drive a couple of days ago, when you first suggested it. Should I do it again, after the Safe Boot? Should I delete the index on the source drive?

No, if it didn't help the first time, that's apparently not the problem. I thought you'd done it, but wasn't sure.

I'd be willing to try deleting the File System Event Store...if that's not too dangerous. 🙂

It's a pretty long shot, but worth a try. You want as little running as possible: the relatively-easy way is to power off or Restart your Mac (don't have it reopen windows), log on as an Admin user, quit all apps you can, disconnect from the Internet if possible.

Start the Terminal app (in your Applications/Utilities folder).

In Terminal, the prompt looks like this: <Computer Name>:~ <your name>$

(where <your name> is your short user name). It's followed by a non-blinking block cursor.

Copy the following after the prompt:

sudo rm -rf /.fseventsd

and press Return. You'll be prompted for your Admin password (which won't be displayed). Press Return again.

There may or may not be any messages displayed; you'll know it's done when you see the prompt again. It shouldn't take more than a few seconds -- a minute or two at the most.

Quit Terminal and start your normal apps, etc., and run a backup. The first one won't be able to use the Event Store, of course, so will have to do a deep scan (with messages in the log), which will take quite a while to compare everything on your Mac to the most recent backup.

Wait a few minutes, perhaps do a few other things, run another backup, and see what happens.

Drat. 😟

Very strange, but a bit of a clue, anyhow -- it does seem linked, to some extent, to the Event Store. There must be something running on your Mac that's causing some sort of problem for it.

Check your .plist files again -- to be sure you don't have another load of the funky ones.

Do you have any anti-virus apps, or file/directory monitor apps, or anything else that might be involved?

Check your logs, too -- look for any messages that are repeated over and over, especially anything with dire-sounding words like failed, exited abnormally, exited with code xx, I/O error, etc.

Barry Lyga wrote:

. . .

One thing: I'm wondering if maybe Crashplan is causing some kind of conflict.

It certainly shouldn't -- and if it did, you'd think it would have been reported by now.

2/17/13 8:46:11.549 PM fseventsd[47]: read_disk_log: problems with history file: /.fseventsd/0000000003b32a37 (fd -1, errno 2 ; is_reg 0; uid 0/0 gid 0/0)

That is interesting. I don't know much about the workings of it to know what that really means. It might be just that the log was replaced, and it now can't find a previous entry that was deleted, or it might be a symptom of something else causing trouble with it.

2/17/13 8:42:12.701 PM SyncServer[3599]: [0x7ff59041b9d0] |DataManager|Warning| Client com.apple.Mail sync alert tool path /System/Library/Frameworks/Message.framework/Resources/MailSync does not exist.

Doesn't seem to be pertinent to TM.

But the weird thing is that as I was typing this response, TM just did an 11-second backup. ("No post-back up thinning needed: no expired backups exist") It's almost random.

Very odd. About all I can suggest is keeping an eye on it, especially if you get more of the long backups, and see if there are any more of the fseventsd messages. If you do, the timing of them vs. backups may contain a clue.

We're pretty deep in the weeds here. 😟

Barry Lyga wrote:

Watching the Console during a backup and it threw 7 of those fseventsd errors in a row while stuck at "backing up 85.9MB of 134.6MB"

Aha! Bigger clue!

That "feventsd" -- could it be related to deleting the File System Event Store?

Yup. fseventsd is the OSX process that handles the event store.

Just what's wrong, much less what may be causing it, is a mystery so far.

Keep your eye on it, and keep us posted.

Anyway, I'll try disabling Crashplan for a couple of days and keep track of how long the backups take.

Worth a shot, but seems doubtful that's the issue.

Bah !Yes Time Machine says what it does,it takes time to do what you want it to do.

if the Time Machine informed us that what we did was to be backed up then it would need time to do it when we do not use the computer if you use it the Time Machine would have to change everytime the user added changes and so it needs a time by itself to reckeon all the changes of the day and log them in the Machine.

Time Machine was forever backing up all day I wanted to shut down for this and that reason but I would be just making more work for the Time Machine a day or so in the future then with Vilefault it did not back up so I stopped Time Machine Then my USB external HD broke so I am now looking at that 997.87 GB available space

to dump my Back up's Back up 's Back up because even a USB external HD can go.

So now I tried Podini's

Or, there is a workaround, to "reserve" some space there, by creating a disk image of the desired size and storing your data there, but it's rather cumbersome, and two Macs can't share it at the same time.

Do not get me wrong, it seemed a good alternative at trhe TimeMachine ,but I set up a whopping 500 GB " reserve " disk image and after 12 hours it was still a sluggish 34 GB so how much Time (Machine) is 500 GB going to take ? How long does it take to erase a 500GB HDD 35 times zero over. 3 days !

So slow but sure !, there is always something in the system that is not right and Time Machine will perform slowly especially with an external HHD to boot

All in all Back up is better done on I very solid external HDD and a potable HDD. From what I have read Time Machine Airport Disk is not solid.

Rajapintaja wrote:

Bah !Yes Time Machine says what it does,it takes time to do what you want it to do.

if the Time Machine informed us that what we did was to be backed up then it would need time to do it when we do not use the computer if you use it the Time Machine would have to change everytime the user added changes and so it needs a time by itself to reckeon all the changes of the day and log them in the Machine.

Actually it normally does takes very little time to figure out what's changed and needs to be backed-up. And normally it runs just fine while you're using your Mac -- if it doesn't, then something is wrong with your setup, hardware, etc.

Time Machine was forever backing up all day

Again, that's not normal. Something is wrong.

then with Vilefault it did not back up so I stopped Time Machine

There are two versions of FileVault. On Leopard and Snow Leopard, it encrypted your home folder into a sparse bundle disk image, and only backed-that up when you were logged out. Not a good combination with Time Machine.

On Lion and later, File Vault 2 is an entirely different thing -- your entire disk is encrypted, and Time Machine runs normally -- meaning it backs-up whatever's changed hourly, if you let it. That's normally quick and usually not noticed at all.

Then my USB external HD broke

That could explain why your backups were slow -- often as a HD is failing, it will get slower and slower before finally breaking.

So now I tried Podini's

Or, there is a workaround, to "reserve" some space there, by creating a disk image of the desired size and storing your data there, but it's rather cumbersome, and two Macs can't share it at the same time.

That applies only to backing-up to a Time Capsule's internal HD, when you have other stuff on it (not recommended).

All in all Back up is better done on I very solid external HDD and a potable HDD.

Correct.

From what I have read Time Machine Airport Disk is not solid.

If you mean a USB drive connected to an Airport Extreme, that's usually unreliable and is not supported by Apple. If you mean a USB drive connected to a Time Capsule, that is reasonably reliable and is supported. Of course, anything done over a network will be slower, especially if done wirelessly.

Please clarify what you're backing-up to, what version of OSX you're running, and what your current problem is.