Question

My time machine is backup is very slow. Connection is FireWire 800. Transfer is around 5 Mb/s or 0.5 MB/s, which seems slow. I've seen plenty of mention online of very slow Time Machine backups. What are some techniques or suggestions to speed this up?

I have:

  • Run the Disk Utility Repair tool. Everything is fine.
  • Disabled Spotlight on the remote disk

Searching for backupd in system.log gives the following:

13:29:33 Starting standard backup
13:29:33 Backing up to: /Volumes/BACKUP_DISK/Backups.backupdb
13:29:36 Event store UUIDs don't match for volume: Macintosh HD
13:29:36 Waiting for index to be ready (101)
13:29:55 Node requires deep traversal:/ reason:must scan subdirs|new event db|
13:42:53 No pre-backup thinning needed: 39.84 GB requested (including padding), 71.68 GB available
13:42:53 Waiting for index to be ready (100)
14:23:37 Bulk setting Spotlight attributes failed.
14:29:39 Copied 2.0 GB of 28.7 GB, 69560 of 113624 items
15:29:57 Copied 2.8 GB of 28.7 GB, 69560 of 113624 items
16:30:00 Copied 3.3 GB of 28.7 GB, 69560 of 113624 items
17:30:04 Copied 3.8 GB of 28.7 GB, 69560 of 113624 items
18:30:07 Copied 4.3 GB of 28.7 GB, 69560 of 113624 items
18:51:23 Bulk setting Spotlight attributes failed.
19:30:07 Copied 4.8 GB of 28.7 GB, 69999 of 113624 items
20:30:19 Copied 5.4 GB of 28.7 GB, 69999 of 113624 items
21:30:28 Copied 6.2 GB of 28.7 GB, 69999 of 113624 items
21:38:36 Stopping backupd to allow ejection of backup destination disk!
Was it helpful?

Solution 2

Unfortunately, the solution in this case was:

Format disk (lose old backups)

Now the disk backs up at full speed.

OTHER TIPS

On newer OS - the io is throttled in some cases. You can review these steps in case they apply in your case.

Time Machine ridiculously slow after El Capitan upgrade

The throttle should be evident in observing the backupd usage of the filesystem:

sudo fs_usage backups

Writing a command to /etc/sysctl.conf and then restarting should disable that throttle:

echo 'debug.lowpri_throttle_enabled=0' | sudo tee -a /etc/sysctl.conf

You'll want to watch for other performance issues (extra CPU usage, other IO slowing, increased energy usage when backing up) after making that low level system tuning adjustment.

According to this thread:

Delete the "in progress" folder on the TM drive and then things will operate at a normal speed.

Note: That did not fix the problem in my case.

You could try renaming the /.Spotlight-V100 folder on the TM volume. I've never tried this but Spotlight is used to decide which files are new/changed since the last backup. The only thing that comes to mind in reading your question, is maybe the index is broken and needs to be rebuilt, and TM in the mean time is running that search live, rather than indexed.

Long shot.

Add the Time Machine disk to the Privacy list of Spotlight (in System Preferences) so that Spotlight does not spend time accessing that drive.

Fragmentation

In my experience, it's very common for the catalog B-tree to become very fragmented. On a volume that is both large and full, this type of fragmentation can have a significant effect.

At any time

Use the demo version of iDefrag to view statistics.

When the backup volume is as close as possible to full

Consider using the Metadata algorithm of the full version of Defrag, or a comparable utility, to defragment the catalog and other critical files.

(For a Time Machine backup volume, don't bother with full defragmentation.)

Other utilities

fileXray — this alone can not defragment the relevant files, but for analysis: I know of nothing better than fileXray.

Cross reference

Do Macs need to be defragmented?

It could also be related to MRT process. For me, when I removed MRT process, backup speed becomes normal. By the way, MRT is sth called malware remove tool which comes with the security update and it will take lots of CPU to do scanning work. There are some articles on how to remove it.

Licensed under: CC-BY-SA with attribution
Not affiliated with apple.stackexchange
scroll top