Q: Time Machine Frequent "Start New Backup"

aschmid wrote:

 

Okay my bad. It's not "consolidating" it's "thinning" by keeping the first and deleting the rest. I think the point was more on that's actually deleting something not so much on what and how exactly.

You were suspecting a link between rssalling's roughly monthly verification failure with the thinning, but it's not connected to thinning, since thinning is done as part of every backup. 

The automtatic verification is only run monthly -- so if whatever goes wrong isn't detected by the actual backup, or a manual Verify Backups, it will probably only be detected monthly.  There's really no way to tell when the damage actually occurred.

It's usually caused by wireless or power problems, or a NAS that's not fully compatible with the version of OSX.

See #C13 in Time Machine - Troubleshooting, especially the pink box.

Message was edited by: Pondini

Here is two examples of backups, one that does thinning the other one that doesn't:

Aug 10 12:43:28 iMac27 com.apple.backupd[2785] <Notice>: Starting post-backup thinning

Aug 10 12:43:57 iMac27 com.apple.backupd[2785] <Notice>: Deleted /Volumes/Time Machine Backups/Backups.backupdb/iMac27/2012-07-08-083336 (1.09 GB)

Aug 10 12:44:11 iMac27 kernel[0] <Debug>: hfs: cat_resolvelink: can't find iNode1435250

Aug 10 12:44:14 iMac27 kernel[0] <Debug>: hfs: cat_delete() failed to delete thread record on volume Time Machine Backups

Aug 10 12:44:14 iMac27 kernel[0] <Error>:

Aug 10 12:44:14 iMac27 kernel[0] <Debug>: hfs: Runtime corruption detected on Time Machine Backups, fsck will be forced on next mount.

This one doesn't do thinning:

Aug 23 01:46:30 iMac27 com.apple.backupd[15540] <Notice>: Starting post-backup thinning

Aug 23 01:46:30 iMac27 com.apple.backupd[15540] <Notice>: No post-back up thinning needed: no expired backups exist

Aug 23 01:46:30 iMac27 com.apple.backupd[15540] <Notice>: Backup completed successfully.

I don't want to argue about the internal of Time Machine, all I am saying is that I have reduced my failure rate on a Synology NAS box by switching from a "timemachine" user to the admin user when doing backups.

aschmid wrote:

. . .

Aug 23 01:46:30 iMac27 com.apple.backupd[15540] <Notice>: No post-back up thinning needed: no expired backups exist

Exactly.  On every backup TM looks for expired backups.  If there are any, they'll be deleted.  If backups are very regular, there will be at least one on nearly every backup; if they're mor sporadic, there will be none on some, more than one on others. 

all I am saying is that I have reduced my failure rate on a Synology NAS box by switching from a "timemachine" user to the admin user when doing backups.

That's something you do on the NAS setup, not within Time Machine, right?  The actual backup runs as the root user, no matter what you do on your Mac.  

If that really makes a difference, there's some sort of internal or setup problem with the NAS.  You might want to let them know your findings.

Hello again

Again I had the message saying that Time Machine verification needs to redo the whole backup.

I have since found this article about fixing time machine sparse bundles on NAS

I followed the instructions which took a while to work out what my qnap server was called and why the sparsebundle could not be found but I go there in the end. Timemachine has just finished the verification, prepared the backup and is now backing up again.

For how long I am not sure but it is backing up 1.88Gb as I type.

I upated to the latest beta software on my Synology last week, however I got the message again today.

I never had this under Snow Leopard.

This is getting frustrating.

Pondini wrote:

 

It's usually caused by wireless or power problems, or a NAS that's not fully compatible with the version of OSX.

 

See #C13 in Time Machine - Troubleshooting, especially the pink box.

 

 

that's what you've posted one hundered times.

you have no other suggestions?

Wireless problem? there are lot's of wired NAS users who have that problem, but wait: the NAS could not be fully compatible with TM or OSX, except for those users who use a Time Capsule, but they have all power problems, eh????

This is a Time Machine and / or OSX bug, and at Apple they do not even have clue. That's it. Deal with it.

MacPear wrote:

 

Pondini wrote:

 

It's usually caused by wireless or power problems, or a NAS that's not fully compatible with the version of OSX.

 

See #C13 in Time Machine - Troubleshooting, especially the pink box.

 

 

that's what you've posted one hundered times.

At least, for the very good reason that those are the most common causes, so the logical places to start.

you have no other suggestions?

Sure.  Have you completely ruled out wireless problems?  Both marginal connections and interference, even intermittent ones? Power problems (everything is on UPS systems)?

And NASs aren't perfect, either.

Wireless problem? there are lot's of wired NAS users who have that problem, but wait: the NAS could not be fully compatible with TM or OSX, except for those users who use a Time Capsule, but they have all power problems, eh????

Yes, there are lots of possible causes.  

This is a Time Machine and / or OSX bug, and at Apple they do not even have clue. That's it. Deal with it.

Really?  You jump to that conclusion based on a few incomplete reports?  Or have you, for example, tried other backup apps in the same circumstances (hourly backups over the same configuration for several months, with a verification procedure at least monthly)?

By all means, offer your technical expertise to Apple: http://www.apple.com/feedback/ 

Or, better, file a detailed Bug Report, per Reporting a Problem to Apple.

Message was edited by: Pondini

Can we please stop blaming wireless for this problem. Time Machine must work over wireless - if it can't it is not fit for purpose and needs re-designing. Even if wireless problems can be shown to be the immediate cause, the real problem is with Time Machine's inablility to deal correctly with network problems. Enough people are experiencing this to indicate a problem and it needs fixing - by Apple.

Wireless is just another computer networking technology and all computer networks are unreliable. Fortunately, the pioneers of networking were smart enough to realise this and gave us TCP, which provides a reliable service over an unreliable, best effort, network (IP). Wireless is no different.

Time Machine must be designed to cope with network interruptions and roll back to the last good state in a safe manner. Trashing the entire backup history every few weeks (in my case) is just not acceptable. No excuses.

I was hoping that the latest version of Netatalk incorporated in Synology DSM 4.1 would provide a fix, but sadly it seems from what others have posted that it won't.

ritec wrote:

 

Can we please stop blaming wireless for this problem. Time Machine must work over wireless - if it can't it is not fit for purpose and needs re-designing.

Yes, it should.  In theory, HFS+ journalling ought to prevent damage, and in most cases, it does.  But, like most such things, it doesn't always. 

And that's not limited to wireless -- pull the plug while copying or backing-up to a directly-connected external HD formatted HFS+.  Most of the time, the damage that causes can be fixed by Disk Utility, but not always.  Sometimes it can be fixed by 3rd-party apps like DiskWarrior, but again, not always. 

Nobody is saying wireless is always the cause; it just seems to be the most common.

But, you're not talking to Apple here -- this is a user-to-user forum. 

By all means, file a detailed Bug Report, per Reporting a Problem to Apple.

I have the same problem but my girlfriend has the same MacBook Pro (only mid 2011, mine is begin 2011) with excaclty the same version of Lion and we both use the same WD Mybook World NAS to make Time machine backups. On her computer this runs without any problems for months and still works, but I have get the backup verification error all the time.

Just thought I would chime in also.

I am having the same issue with a MacBook Late 2009 (Aluminum) running 10.8.1.

So if it was a Lion thing it has carried over.

I am runnning a Synology DS409, I'm still on Version 3.X as I dont need the features of 4.1 and hey, not broke dont fix. I have also followed threads that say that people are still having the issue on DSM4.1.

Of course its backing up over 802.11n because, well that’s the point of it isn't it.

My base station is pretty close, so it's not signal strenght. Again even if it was, thats not the point!

Yes, it should.  In theory, HFS+ journalling ought to prevent damage, and in most cases, it does.  But, like most such things, it doesn't always ...

That is not the point.   Any backup instance can fail for any of the reasons - network problems, power failure,   computer went to sleep,  cat tripped over power cord,  whatever.   But a minimum requirement of any backup system worth its name (and there are few around) that under no circumstances, barring disk failure,  failed or corrupted backup instance should have an ability to invalidate previous backups.  It is just a design flaw.    And of course a backup that you may need to start anew without notice is not a backup at all.    Imaging getting an e-mail from your cloud backup - "sorry, you last backup failed due to network problems and we have to erase all your history" - what would you reply ?

dmpogo wrote:

. . .

Any backup instance can fail for any of the reasons - network problems, power failure,   computer went to sleep,  cat tripped over power cord,  whatever.

Exactly.

But a minimum requirement of any backup system worth its name (and there are few around) that under no circumstances, barring disk failure,  failed or corrupted backup instance should have an ability to invalidate previous backups. 

Surely you're not saying that only happens with Time Machine?   I've had it happen with CarbonCopyCloner (which is a great product, but, like all other backup apps, is not perfect).  And that was on a directly-connected F/W 800 drive, not a wireless network backup.

It is just a design flaw.

Call it whatever you like, it's a fact -- no backup app is perfect.   That's why many experienced users advise never to trust your data to a single backup, no matter where it is or how it was made. 

Imaging getting an e-mail from your cloud backup - "sorry, you last backup failed due to network problems and we have to erase all your history" - what would you reply ?

It happens.  Read some of the blogs about Mozy, Carbonite, Daisy Disk, etc.

But the bottom lines remain:

If you're so sure it's only a Time Machine problem, then by all means use a better app, if you can find one.

Occasional corruption of network TM backups, especially wirelessly, has has been going on, to one degree or another, since Time Machine was released (with Leopard, about 5 years ago). The automatic verification that detects it and produces the message started with 10.6.4, over two years ago. 

Venting here does not get Apple's attention; this is a user-to-user forum -- Apple does not monitor it for technical content. File a detailed Bug Report, per Reporting a Problem to Apple.

Hi Pondini

I think the reason he is here is for answers. I know I am.

I only get this on one machine (as meany people do) so I dont think it’s a Time Machine issue, I think there is a bug somewhere

As Bart85 says above, he has 2 macs, one works one does not.

There must be a solution. I have raise the issue with Synology and am waiting to hear back.

You may look at Mountain Lion.   If I am not mistaken it allows Time Machine to perform backups while laptop is sleeping, which may make failures less frequent.   I don't think Synology per se has much to do with the situation - I got an error backing up my wife Macbook Air over wifi to the homebrew Linux server with exports one disk with netatalk.   Time machine ran fine for 3 months and few days ago threw this error.    The issue I think is in design - as a backup space Time Machine tries to manage its own filesystem within one sparse file, and doing this over remote connection is unreliable in the long run. And once filesystem is corrupted, the whole backup history is lost. So if it happened once, it will be happening periodically.   I left TimeMachine to run on my wife's Air as a short term, user friendly interface to few recent copies, but switched to Linux server side Backuppc as a long term backup solution (it just initiates rsync transfers over ssh from MacBook Air according to the schedule of your choice).