• Time Machine unbelievably slow


    rctneil
    • Minor
    Message added by limetech

    Please be aware that these comments were copied here from another source and that the date and time shown for each comment may not be accurate.

    Hello,

    I have a Time Machine share set up on my server to backup my MacBook Pro too but it's unbelievably slow.  Any suggestions how to speed it up or figure out why it's slow?

     

    I'm using a 2017 13" MacBook Pro running Catalina 10.15.2. Connected via Ethernet as well.

     

    All suggestions very much welcome!

     

    Thanks,

    Neil

     

    Edit: As per @limetech instructions I have attached diagnostics:

     

     

    tardis-diagnostics-20200204-2055.zip

    • Like 1



    User Feedback

    Recommended Comments



    49 minutes ago, amiskell said:

    Did you add the proper configuration to avahi in order to advertise the shares?

    Yes, I set it up exactly like my bare metal Ubuntu 19.10 setup, which is currently working perfectly for the last several days. 

    Link to comment

    I am using rsync from my Ubuntu box to put the backups on UNRAID and getting between 80 - 120 Mbps transfer speeds. 

     

    Something is weird with MacOS + UNRAID + SMB I guess. 

    Link to comment

    Found another thread that might be relevant.  Updated all my TM backups to be on the disk share instead of the user share, disabled hardlinks, and set Case sensitive to yes.  Seems like SMB has trouble with folders that have thousands of files, like a Sparsebundle.

     

    My CrashPlan docker that backs up the TM share is running a lot faster.  Waiting to see how TM backups run.

     

     

    • Like 1
    Link to comment

    Switching Case-Sensitive Names to 'Yes' on my Time Machine user shares seems to have solved most of it here. My large (~2.6 TB) backups are completing normally again. No other changes.

     

    Unraid 6.8.3 + macOS Catalina 10.15.6 (just updated, same in Mojave)

    • Like 2
    Link to comment

    Seems like i'm still having issues, the other day I let a large backup run overnight and it backed up something like 8 GB over 10 hours.  

     

    Something that may or may not be relevant is in the Open Files plugin I'm seeing there's always a few dozen files listed as being open from each of my BackupBundles, and in the Main Tab, I can  often see the disk being accessed(in the few hundred KB/s read/write range) even when Time Machine isn't currently running.  Opened a ticket with Apple Support, I'm wondering if some of the performance issue is the disk thrashing due to both my Macs accessing it at the same time.

    Link to comment

    I've tried all the solutions in this thread and my backups are painfully slow.  I'm at a loss now of what to do.  My last full backup to TM was over a month ago now because it takes sooooo long.

     

    @limetech is there a full solution? We need to be able to reply on backups completing.

    Link to comment

    Wow, I thought I was only me. I started deleting the in progress backups and all that. It's stupid slow over 1gb link. Plus I'm using cache drives in case it needs to writes new files. 

     

    Based on what people say, I'll stop using time Machine all together for the time being. Please fix by 6.9.x pleaaseee

     

    Link to comment

    I'm not sure this is specific to unraid. I ran across this thread because I'm having the same issues backing up to a shared drive connected to another Mac. I ended up setting the backup to happen once a day starting when I go to bed since it sometimes takes hours just to free up disk space or to back up more than 1GB of data. About the only time it works properly is if it has very little to back up and doesn't need to trim old backups. I'm just going to switch back to using a USB hard drive since Catalina and Time Machine over the network is too unreliable and slow. 

    Link to comment
    On 9/13/2020 at 10:46 AM, AJ08 said:

    I'm just going to switch back to using a USB hard drive since Catalina and Time Machine over the network is too unreliable and slow. 

    In my case, it was definitely specific to Unraid 6.8.x, even in Mojave. I had held off upgrading both for a while and did the Unraid 6.8 upgrade first.

    Link to comment
    On 8/25/2020 at 1:57 PM, KernelG said:

    Switching Case-Sensitive Names to 'Yes' on my Time Machine user shares seems to have solved most of it here. My large (~2.6 TB) backups are completing normally again. No other changes.

     

    Unraid 6.8.3 + macOS Catalina 10.15.6 (just updated, same in Mojave)

    This totally seemed to have fixed it for me as well. Over 300GB in a few hours, where earlier I would barely see a few GB before it seemingly crapped out.

    Thanks for the tip!

    • Like 1
    Link to comment

    I'm still having issues with this on 6.9.1.  Tried the case sensitive trick and starting fresh with a new share, but no luck.  Started a new 100GB backup last night around 12am and it only did about 75GB by 8am.

    Edited by acosmichippo
    • Like 1
    Link to comment
    3 hours ago, acosmichippo said:

    I'm still having issues with this on 6.9.1.  Tried the case sensitive trick and starting fresh with a new share, but no luck.  Started a new 100GB backup last night around 12am and it only did about 75GB by 8am.

     

    I noticed that too! This thing is driving me crazy, I have tried everything I found online and on this forum with no success.

    Link to comment

    well today i tried again. I read somewhere that the TM sparesbundle really doesn't play well with multiple disks, so  I made another brand new TM share using all the same suggestions as before, but I also set the share to use only one disk in the array.  That seems to be working pretty well, TM currently estimates a 103GB backup will take 3 hours which is a TON faster than my last attempt of ~75GB in 8 hours.  (Also this is over wireless, so I imagine it would go even faster if I was wired.)

     

    I'll keep an eye on how subsequent backups do since I noticed some people say they can go back to a crawl.

     

    Coincidentally I have also recently upgraded my SSD cache, so I'll probably just enable the cache on the TM share once this troubleshooting is done.

    Edited by acosmichippo
    Link to comment
    5 hours ago, acosmichippo said:

    well today i tried again. I read somewhere that the TM sparesbundle really doesn't play well with multiple disks, so  I made another brand new TM share using all the same suggestions as before, but I also set the share to use only one disk in the array.  That seems to be working pretty well, TM currently estimates a 103GB backup will take 3 hours which is a TON faster than my last attempt of ~75GB in 8 hours.  (Also this is over wireless, so I imagine it would go even faster if I was wired.)

     

    I'll keep an eye on how subsequent backups do since I noticed some people say they can go back to a crawl.

     

    Coincidentally I have also recently upgraded my SSD cache, so I'll probably just enable the cache on the TM share once this troubleshooting is done.


    My Time Machine share is already on one drive (excluding parity), so I don't think that's the reason for your speed bump.

     

    I don't know why, and maybe this is your case, the first Time Machine backup, on a brand new share, is always fast and without slowdowns.

     

    Over time I could also learn how Time Machine does not play well with cached shares, before 6.9.1 I had these problems with the cache enabled, once I removed it, backups became fast.

     

    Now with 6.9.1 we are in the same situation again, even without cache.

     

    However keep us updated on the results of the subsequent backups, with and without cache.

    Edited by Gian Marco Cinalli
    Link to comment
    On 3/29/2021 at 4:02 PM, acosmichippo said:

    good point, I should have realized that the initial backups would usually be on a single disk anyway.

     

    Hi, do you have any updates on this?

    Link to comment

    It's hard to say because I haven't tested it on a wired connection.  Subsequent partial backups still seem really slow though.  Like right now it's backup up 1.5GB and calculates 2 hours remaining.   I would think even on wifi it should be faster, but maybe I'm in a dead spot.

    Link to comment
    On 4/20/2021 at 3:10 PM, acosmichippo said:

    It's hard to say because I haven't tested it on a wired connection.  Subsequent partial backups still seem really slow though.  Like right now it's backup up 1.5GB and calculates 2 hours remaining.   I would think even on wifi it should be faster, but maybe I'm in a dead spot.

     

    I'm on a wired connection and I can confirm that is unusable.

    Link to comment

    Found this thread while searching for the same issue I have been fighting with the two MacBooks in the household, figured I would add things I have tried.  One is always wired with 1GB connection and the other is always on WiFi - both running Big Sur.  I'm using TimeMachineEditor to limit the backup to network 2x a day to minimize the drives being spun up.  I've also noticed the "bursty" networking mentioned earlier during a backup, and if I have a 1GB backup - it'll take upwards of 20-30 minutes.  Even setting it to backup to Cache, then run mover later is painfully slow.

     

    The kicker is if I try to copy over files using SMB and Finder - to either a normal share or even a "Time Machine" share - speeds are great and consistent.  In my research I've also seen folks having similar issues on Synology devices too.  I did come across this page https://osxdaily.com/2016/04/17/speed-up-time-machine-by-removing-low-process-priority-throttling/ which has a terminal command to turn off "throttling" of the time machine backups.  When setting it to "0" there is an extreme increase in speed for time machine, especially on an initial backup.  But after setting back to "1" my "incremental" backups are still very slow, as it seems leaving it on "0" isn't necessarily recommended.

     

    I also have a few Windows devices which backup using Veeam beautifully - but Veeam for Mac support is in its infancy, and also doesn't support full system restores.  I have been trialing Acronis True Image on the Macs - pointed at a standard SMB share on Unraid for the past week and it has been working great.  The initial (250GB) backup only took about 45mins and the 2x daily incremental are finishing in 10 mins or so.  I hate to have to use another 3rd party software to do backups when Time Machine is built to do the job, but it feels like Apple just wants you to pay for iCloud storage and use that.  If Acronis works out, there's usually a sale on it for $40 / 3 devices or less so I will just end up going that route.

     

    Coupled with the copy tests, and other NAS device users also experiencing issues, it makes me wonder if it's not so much a "Mac/Unraid" thing as it is Time Machine "working as designed".

    Edited by WackyWRZ
    Link to comment

    With the new option of having multiple pools, i set up a pool with a single HDD and set my Time Machine share to HDDPool only.  This seems to have solved the performance issues for me, though i've also said that before and had the speeds drop over time so I'll try to keep an eye on it over the next few weeks.  I wonder if there's something about the way sparse bundles are managed that causes TM to do a lot of random reads/writes that trashes performance on things like arrays, particularly with simultaneous access as opposed to having a single dedicated disk as a target like with the Time Capsule.  Seems like people tend to get good speeds for initial backups and they drop off over time, so maybe fragmentation of the sparsebundle is an issue too?

    • Like 1
    Link to comment

    My Time Machine backups to my Unraid server were not working reliably (slow speed, backups never completed). I followed the suggestion to turn off hard link support and I created a fresh backup on a new share (only one assigned disk). After 3 days of testing, backups are completing reliably.

     

    If anyone else comes across this thread, I recommend this method. I'll report back if anything changes.

     

    Edited by phunktional
    Link to comment
    14 minutes ago, phunktional said:

    My Time Machine backups to my Unraid server were not working reliably (slow speed, backups never completed). I followed the suggestion to turn off hard link support and I created a fresh backup on a new share (only one assigned disk). After 3 days of testing, backups are completing reliably.

     

    If anyone else comes across this thread, I recommend this method. I'll report back if anything changes.

     


    Hi, your share is in the array or is a separate pool of only one disk?

    Link to comment



    Join the conversation

    You can post now and register later. If you have an account, sign in now to post with your account.
    Note: Your post will require moderator approval before it will be visible.

    Guest
    Add a comment...

    ×   Pasted as rich text.   Restore formatting

      Only 75 emoji are allowed.

    ×   Your link has been automatically embedded.   Display as a link instead

    ×   Your previous content has been restored.   Clear editor

    ×   You cannot paste images directly. Upload or insert images from URL.


  • Status Definitions

     

    Open = Under consideration.

     

    Solved = The issue has been resolved.

     

    Solved version = The issue has been resolved in the indicated release version.

     

    Closed = Feedback or opinion better posted on our forum for discussion. Also for reports we cannot reproduce or need more information. In this case just add a comment and we will review it again.

     

    Retest = Please retest in latest release.


    Priority Definitions

     

    Minor = Something not working correctly.

     

    Urgent = Server crash, data loss, or other showstopper.

     

    Annoyance = Doesn't affect functionality but should be fixed.

     

    Other = Announcement or other non-issue.