ruggiero

Q: Time Machine Backups to Server fail

I am having a weird but grave problem.

 

Setup:

DATASTORE: MacMini (Late 2014), Core i5, 1.4GHz, 16GB RAM, Yosemite 10.10.5 with Server.app 5.0.15, 32TB in a ThunderBay4 cabinet. Everything brand new (1 month old), no reported errors whatsoever (SMART, hardware diagnostics, etc). Had this machine running with el Capitan, too. Same configuration, same issues (s. below)

 

MY_SERVER: MacMini (Mid 2011), Core i5, 2.3GHz, 8GB RAM, OS X 10.9.5, Server 3.2.2

 

SOME_CLIENTs: Several MacBooks (pro) with el Capitan Clients, some with, most without Server.app

 

DATASTORE is set up to provide Time Machine service to the other systems. There are several partitions on those 32TB that are served as TM locations, so different clients can have different quota for their backups as max backup size can only be set per location, not per individual client. :-(

 

MY_SERVER and the various clients all use DATASTORE as TM destination. All the backups fail more or less regularly with weird error messages like Backup failed with error=(null). Sometimes however backups work for a couple days, then fail again. Sometimes deleting the backup DMG and starting a new backup fixes things.

Interesting fact is that MY_SERVER has been backing up reliably, no hick-ups for several weeks (since the DATASTORE machine was being put into service)

 

My God, just now, while writing this, I had a failing backup on MY_SERVER. Console.app on MY_SERVER shows these errors:

 

19. 8. 16 10:44:31.000 com.apple.backupd[93194]: Failed to eject volume /Volumes/tm_miniserver (FSVolumeRefNum: -704; status: -47; dissenting pid: 0)

19. 8. 16 10:44:31.001 com.apple.backupd[93194]: Giving up after 3 retries.

19. 8. 16 10:44:31.006 com.apple.backupd[93194]: Backup failed with error 21: 21

 

To really clean up things I rebooted MY_SERVER and started a backup manually right after reboot. Now backup fails with yet another nice message: The backup on "datastore" is already in use. And console this time notes

 

19. 8. 16 11:04:47.663 com.apple.backupd[1104]: Backup failed with error 31: 31

 

What's that now? Corrupt backup disk image? Somehow locked? Server.app on DATASTORE reports "Backup in progress". How can I fix this?

 

Each of the clients has a local USB disk that is also used for backup, so the clients are covered. But why is Time Machine to DATASTORE such a fail? Tons of problems.

 

What I have seen is that CPU load on DATASTORE is rather high, Apple File Service consuming most of the cycles. Could it be that the MacMini is too slow?

 

Thoroughly confused and at my wits end.

MacBook Pro with Retina display, OS X El Capitan (10.11.6)

Posted on Aug 19, 2016 2:33 AM

Close

Q: Time Machine Backups to Server fail

  • All replies
  • Helpful answers

  • by tim.rohrer,

    tim.rohrer tim.rohrer Aug 19, 2016 8:40 PM in response to ruggiero
    Level 1 (36 points)
    Servers Enterprise
    Aug 19, 2016 8:40 PM in response to ruggiero

    Ruggiero--

     

    I've got the same errors on one remaining laptop that I cannot seem to clear. Some aspects of my situation appear to be different, but let me share them and see if we can find more commonality.

     

    My setup:

    OS X 10.11.6/Server 5.1.7 on a 2014 Mac Mini. SOHO network with three workstations using Time Machine service for backups, configured by Profile Manager. Three workstations are: 1) 2013 Mac Pro, 2) Macbook Pro, 3) Macbook. Until a few months ago, all workstations were Yosemite.

     

    A few months ago, I upgraded the Mac Pro to El Capitan. At that point, backups from that computer broke. What I discovered by watching the file sharing connections what that many (many many!) connections would get established. File Sharing on the server showed multiple connections from the user, some growing stale. On the workstation, multiple mounts would get created and usually remain, with the -x number increasing. For example: /Volumes/<TMB Share Name>-1. On the workstation, the user would get an error that the Time Machine Backup failed because of the backup was in use. I worked on this for several days, finally getting it fixed by removing the Time Machine profile in profile manager, deleting the backup for that workstation, and reconfiguring it. During part of the reconfiguration, I logged into the workstation with an admin account and deleted /Library/Preferences/com.apple.TimeMachine.plist. I *may* have actually removed all the profiles from the workstation, but I don't think so. In any case, it fixed the problem and so I was happy.

     

    Last week, I upgraded the two laptops to El Capitan, and guess what. Backups for each started failing with the same symptoms. For the Macbook Pro, following the steps above seems to have worked, and now that it is reconfigured, backups have been working as they should.

     

    However, I'm not getting these steps to work with Macbook. Although I'm not getting countless AFP connections, I'm getting /Volumes/<TMB Share Name>-1, and the error 31 in the logs.

     

    Next, before I call it a night, I'm going to go through the steps again and also check the laptop's /Volumes to make sure nothing is left. I'm noticing /Volumes/<TMB Share Name> is still present even the logs show the backup failed. It appears something is not getting cleaned up.

     

    I'll post an update in the morning, but I'm curious if you're using Profile Manager to configure the clients and MY_SERVER for the backups?

  • by tim.rohrer,

    tim.rohrer tim.rohrer Aug 20, 2016 3:27 PM in response to tim.rohrer
    Level 1 (36 points)
    Servers Enterprise
    Aug 20, 2016 3:27 PM in response to tim.rohrer

    It appears the actions I described worked. For some reason, I guess I didn't have things clean enough and had to do it again.