Time Machine Backup Stalls after El Capitan Upgrade

You can save the backups to another drive, but keep in mind that they are corrupted, so you may or may not be able to restore from the old backups without risking corruption to the files on your Mac.

Unless you really need the old backups, it would be safer to let Time Machine erase the old backups and start another backup set.

Having the exact same issue. 2015 MacBook Pro 13". Never an issue with Time Machine. After upgrading to El Capitan 10.11.1 I suddenly got the "must create a new backup for you". Annoying, but fine. I copied the sparse bundle to a USB disk and deleted it, then connected my computer via a Thunderbolt->Gigabit adapter and let Time Machine backup again. It stalled at about 112GB (out of 300+ GB) or so. No processes taking CPU power. Just a full, dead stop. There are errors in the logs about MDImporter (iirc) related to backupd - so it seems there is a bug in Spotlight that is causing Time Machine to stop working.

I've tried backing up to a Thunderbolt disk directly connected to the computer, same issue. Tried deleting my Spotlight indexes and re-indexing my drive. Same issue.

Basically, Time Machine is 100% useless on El Capitan 10.11.1 for me. Can't finish a backup. Fortunately everything I have that is critical is backed up to the cloud on a regular basis but the convenience of Time Machine (ability to recover old versions, etc) is missed.

Hoping Apple fixes this in 10.11.2.

Basically, Time Machine is 100% useless on El Capitan 10.11.1 for me

Upgrade install is at least part of the issue.. it should be clean installed.

However if you have 300GB of stuff on the disk a clean install is going to be a lot of work recovering your files.

You did everything right.. using gigabit to a TC.. and then a thunderbolt disk, but it sounds to me like the main disk is corrupted.. especially if it falls over at the same point. The TM logs should tell you what the problem is.

Let me strongly suggest a couple of things.

1. If clean install is not possible at least do a verify of your local disk.. to do this properly you need to boot into recovery. You still should backup beforehand. Which is like a catch22.. hence point 3 below.. step out of the circle.

2. Make sure if you loaded any anti-virus it is turned off. So far anti-virus software on a Mac causes more harm than good.

3. Use a different backup software in the interim.. I have been recommending Carbon Copy Cloner since Mavericks.. as a far more solid usable backup software.. it doesn't have all the niceties of TM but it is a lot more reliable.. Chronosync is also popular.. not one I have used..

So, after about a week of a freshly installed El Capitan OS X on my 2009 Mac Pro, it solved the issue from not even being able to start up Time Machine, but now it's stopped doing automated backups. I have to manually start it, which it will perform completely.

All in the console, I'm getting heaps of SpotLight errors and other issues.

I have a 2013MBP 300GB SSD, the current time machine backup of 17GB has been running for 54 hours and has only completed 1.8 GB. Apple please do something about this.

Apple is not here......we are all users, just like you.

If you have problem, and want some help, start a new post and provide as much detail as possible.

Followup

I ended up on the phone for hours with Apple. Lots of hand offs to supervisors, etc.

They told me to reinstall OS 10.11 from Recovery and try to see if I could run Time Machine on the current set of backups.

It didn't work, so they told me the only thing was start over with Time Machine and the Time Capsule, which I did.

It's been running for a few days and seems to be fine. The past is gone

I did find an interesting article that might explain why we are having these problems. Note that it is not a fix or technical. It sort of just describes the business decisions that have gotten us here: http://www.macobserver.com/tmo/article/the-one-thing-we-need-in-os-x-el-capitan- and-didnt-get

http://joeontech.net/why-i-dont-rely-on-time-machine.html

I have been saying it for a while.. it is just obvious answering posts here for a couple of years now.. that TM reliability is going downhill.. it is far too complex.. and when it decides to wipe itself out there is no way to recover your files.

The author there suggests CCC.. see my post above.

Re: Time Machine Backup Stalls after El Capitan Upgrade

There is a time to call it time on time machine..

A lot of these issues also stem from the fact that Apple insists on continuing their use of HFS+ as the default file system of OS X. I remember back in the Leopard (10.5) days the rumour was, leading up to WWDC, that Apple was going to announce a move to ZFS as their default file system of choice and that Leopard would be the first release where this was the case. It never happened (apparently licensing issues with Sun (now Oracle) derailed the whole project), but remnants of the initial porting work were still in the OS X command line tools (zpool, etc) for a version or two. Apple totally ported other best-in-breed Solaris tech to OS X (dtrace being a good example) so why ZFS didn't make the cut baffles me given how much utility it would offer.

ZFS basically implements "Time Machine" by default in that you can just say "zfs snapshot" and it creates a read-only snapshot of your file system that uses no additional space until you start making changes / deleting files. You can then copy these snapshots to other physical volumes, or just keep them in place if you want to have a "restore to old version" style functionality.

Time Machine as we know it implements everything that was just native / built in to ZFS by using a series of full copies and hard symlinks and all sorts of kludgy, hacky silliness. It's no wonder it regularly explodes and requires full rebuilding.

Incidentally, I was able to solve my problem (above) by:

1) Disable Time Machine.

2) Stopping all Spotlight indexes (mdutil -d -a).

3) Destroying all Spotlight indexes (rm -rf /.Spotlight-V100) manually.

4) Repairing a file system permission issue with Spotlight (removed a bunch of folders owned by a non-existant user in a /private/var/folders subfolder that Spotlight created). No idea how they got there but they were there and the logs were complaining about them causing a fatal error.

5) Re-enabling Spotlight indexes and letting them build fully.

6) Re-running an initial Time Machine backup.

So, Time Machine is dependent on Spotlight and Spotlight breaks easily. Hence, Time Machine breaks easily. Not a great design. 😟

1) Disable Time Machine.

2) Stopping all Spotlight indexes (mdutil -d -a).

3) Destroying all Spotlight indexes (rm -rf /.Spotlight-V100) manually.

4) Repairing a file system permission issue with Spotlight (removed a bunch of folders owned by a non-existant user in a /private/var/folders subfolder that Spotlight created). No idea how they got there but they were there and the logs were complaining about them causing a fatal error.

5) Re-enabling Spotlight indexes and letting them build fully.

6) Re-running an initial Time Machine backup.

Thanks.. that is really useful to have.

After 30 minutes of discussions with various Apple techs, they concluded that the TC is bad. Took it to Apple store, they agreed and ordered a replacement.

Meanwhile, bought a LaCie 2TB drive, TM is working fine. Backed up all 500GB of the Mac in about an hour using the Thunderbolt connection. Working fine now for about 4 days.

New TC due tomorrow.

Just wanted to chime in here, I had an unexpected experience.

I'm using a WesternDigital MyPassportUltra with a new MacbookAir running ElCapitan10.11.1

So first I thought my drive was toast, so I bought a new one. Then again TM slowed to a glacial pace (I went to work several days and left my computer to backup, with only a few GB transferred), so I figured TM itself was broken. I was prepared to use CCC to backup instead and so I left my computer on to backup using CCC (which only took 20min or so for 60GB), went to work that day, and when I returned I found that TM had actually completed the backup that day (approx 300GB). I had forgotten to turn it off, thankfully.

Anyways, the point is that... it seemed like it wasn't going to work at all, but just letting the mac do it's thing for many many hours and DAYS... somehow it worked it's self out. That or running CCC sparked TM into action for some reason?

UPDATE: on reflection, I believe that it may have to do with deleting the TM "Backups.backupdb" folder cleanly and restarting a new backup.... since, using CCC, I used the default settings, which included erasing any files on the destination drive that weren't on the source...

I stumbled onto this topic as my machine is doing the same after i finally decided to upgrade to El Capitain this evening I was unable to locate a resolution? I resolved this issue by removing the failing drive from my Time Machine settings, rebooting and then re-adding it. It re-prompted me for my password and then worked like a charm.

I am wondering if during the upgrade process the keychain somehow got corrupted..?

Hope this helps someone. :-)

Hello Everyone,

Well, since El Capitan upgrade I had huge issues with Time Machine (extremly slow, to the point it was a useless backup system).

I digged all I could to find a solution, and tried almost all I could :

- Formatted backups then restart from scratch

- Destroyed then rebuild all spotlight indexes

- Edited some kernel settings (net.inet.tcp.delayed_ack, debug.lowpri_throttle_enabled, ...)

- Tested backups on another Time Capsule

- Tested in WiFi and cable

- Resetted PRAM and SMC

- And so on ...

None of these changed anything ...

But today I found the solution (at least for me), so I'm sharing it with you (found it randomly to be honest) as I'm pretty sure it might work for some of you guys.

So my issue was due to MenuMeter. Since El Capitan upgrade, I've installed MenuMeter 1.9.1.

This version, when disc monitoring is ON makes hard drives performances to crumble.

So I've disabled it (in fact I completely removed MenuMeter, which is a shame 'cause I love this tool ...) and everything came back normal.

My backups are now all good, as they used to be 🙂

The source : https://github.com/yujitach/MenuMeters/issues/19

Hope this helps some of you (I'm sure many of you are using MenuMeter) !

Best Regards,

LittleBigFox