Unfortunately no one can be told what fun_plug is - you have to see it for yourself.
You are not logged in.
I have been getting "Time Machine disk appears to have changed since last backup" pop up when backing up to my DNA 323. (Thanks to this site for most of getting it working so reliably) This has been investigated in various ways on the Apple discussions boards and other places and seems to stem from Apple now checking the UUID of the Disk. I am none too sure what that is, but it certainly messes up my situation. I have been happily backing up to a sparse disc image on a DNS 323 for years until now. I have struggled with the various workrounds proposed, AppleScripts, editing plists and an app to chose the backup disc, but cannot get them to work without me having to reset things each time which defeats the whole point. Has anyone solved this problem? or have a working workaround, preferably for a non administrator on the Mac and can post a walk through? My unix skills are rudimentary (ditto Perl) but with enough pointers....
Cheers LorenzoG
Offline
Hi Lorenzo,
I setup TimeMachine for my Mac on my DNS323 last week. It's quite simple. Read out this post here and walk through the "manual procedure". You're right. You have to add your UUID to the file. After that the TM works fine. Ignore all hints to with network MAC adress in the spare bundle. Just create a spare-spundle and copy that to the samba share. Than edit the file and go. It's a good idea to change the scheduling process, otherwise your harddisk will spin up and down everytime (4h) your Mac is on the network.
http://www.insanelymac.com/forum/index. … pic=184462
Cheers
Stefan
Offline
Hi Uriens Many Thanks.
OK. Partial success (isn't it always). After a little fiddling I got the method suggested to work, I needed the MAC address in the sparse image filename as suggested at the bottom of page 8 of the post. It then worked and continued to work on its own for a backup or three. I had set it all up in my administrator account but normally work in a non administrator account. I logged out of the Admin account and left the Mac running in my normal account overnight. This morning "The Backup Didc seems to have changed since the last backup" (I forget the exact phrase). Rude words I thought but on logging into my Admin account it backed up with no fuss and without resetting anything or even opening the Time Machine Preferences.
I will try to track down what is going on. I have got the Time Machine buddy widget to track the log files http://web.me.com/pondini/Time_Machine/ … oting.html and have some theories to test; access and permissions on the NAS (they differ between my Admin and my normal accounts), if it works when Admin is logged in but is not the active user, if it needs to be individually set up for each user (probably not as the preferences are universal apart from the excluded files and folders). If anyone has any suggestions they will be gratefully accepted.
To clarify the point, this only occured after I had installed the 10.6.5 update. 10.6.4 and earlier had been working happily by the old method without the plist in the sparse image. The postings on the apple support site http://discussions.info.apple.com/threa … tstart=255 http://discussions.info.apple.com/threa … ;tstart=90 http://discussions.info.apple.com/threa … p;tstart=0 all agree that Apple introduced a check on the UUID of the backup disk with 10.6.5 and that this is the cause of the error popup.
What would be the UUID of a disc image on an ext2 formatted HD on the DNS 323? and how would you determine it?
I suspect that the UUID returned would differ depending on the mount point? of the volume. The Mac seems to see Volumes mounted in a different user account as no access folders.
I will report my progress.
Cheers LorenzoG
Offline
Hi LorenzoG,
You should really look at moving to the most recent version of netatalk (2.1.x) and avahi. With this version, you don't need to worry about pre-creating the sparsebundle as TimeMachine will do it for you while connecting over AFP (not SMB). See the following two posts for more information:
http://dns323.kood.org/forum/viewtopic.php?id=5735
http://dns323.kood.org/forum/viewtopic.php?id=5735
Best,
thebeck
Offline
Thanks thebeck
that looks interesting. If this is the way to go someone should update the Mac osx users page on the wiki with all this.
My DNS 323 is running ffp 0.5, ahavi 0.13, netatalk 2.0.5 and libdaemon 0.13 and I was indeed connecting by samba so it should all be within my very rudimentary unix skills. I hope to have time to give it a go over the weekend.
Cheers LorenzoG
Offline
Hi thebeck
I struggled through it this weekend and have got it working nicely. My thanks to you and to the other posters in the Netatalk 2.1.1 thread
http://dns323.kood.org/forum/viewtopic.php?id=5735 nice one guys.
Some notes; I also found that the -cnidserver command needed to be at the end of the line to get things working and I am finding the log file growing rapidly. Once I am fully happy with things I will try stopping the writing to it, comment out the /ffp/sbin/cnid_metad -l log_info -f /ffp/var/run/cnid_metad.log line in appd.sh and remove the -server command in afpd.conf I am thinking. (Please tell me if I am wrong here) The logs were essential for me to get things working though.
Some pointers;
I got some interesting results when AFP and SMB were set up with different names but were connecting to the same volume. This arrangement seemed slightly unstable and I did not test it much.
The mount for your time machine sparseimage cannot be at the top of the Volume as making the sparseimage fails with a -1 error in TM prefs but was fine in a folder/directory.
I tried creating an account on the DNS323 with a limited quota and doing the AFP mount and Time Machine through that to avoid the sparse image growing to the size of your disk problem but could not get it to work. Insufficient privileges error in TM prefs. I suspect that this may have something to do with having r/w access only to Volume 2 or the Quota mechanism that the DNS 323 uses. I may look at this some more later.
I am finding that the DNS 323 is only showing up as a samba server on the Macs and that the iMac G5 on Leopard happily continues to backup via TM and smb but if you use the go menu in Finder and specify an apf connection then the mount set up in AppleVolumes.default connects no problem. I assume specifying other mounts here with more lines of similar code will give more mount points but again I haven't tried this.
This chunk of code
Code:
<?xml version="1.0" standalone='no'?><!--*-nxml-*-->
<!DOCTYPE service-group SYSTEM "avahi-service.dtd">
<service-group>
<name replace-wildcards="yes">homeCapsule</name>
<service>
<type>_afpovertcp._tcp</type>
<port>548</port>
</service>
</service-group>
needs to go into afpd.service. I regret to say that I struggled with this for hours till the penny dropped. (Is there an emoticon for slapping yourself on the head?)
Best of all this has cured the UUID " the identity of youyr backup disk has changed" problem I started with.
Thanks again
Lorenzo
Offline
Since some time ago after upgrading my OS-X to the lastest version time machine backups started to fail... Apparently this is due to a change in OS-X. I wonder if Apple went through all this trouble to make time machine fail on samba shares to force ppl to buy their hardware for backup?? Whatever their reasons were I think it's stupid not to allow backups to remote samba shares... Specially for something that used to work in earlier OS-X versions... Looks like a downgrade...
Anyways manually creating and copying the sparse image to the NAS solves the problem... until the next time that timemachine decides to fully rebuild your backups... Then you have to go through it again... I'm getting tired of that and I'm looking for other backing up alternatives. If I don't find any satisfactory free/open alternative in a short time I'm probably going for the trouble setting up rsync on cron or something similar...
Offline
Greetings
An Update on my post of 2010-12-20 07:56:59.
Things have continued to work reliably and I am 99% satisfied. 3 (minor) problems remain;
1) All attempts so far to stop the logging to the afpd and cnid logs so far have just made the cnid.db give problems and the connection go into read only but I can live with deleting the log files once in a while.
2) Adding extra mounts lines to AppleVolumes.default works but only (for me) if they have no spaces in the folder/directory name. I have come across a suggestion of how to deal with this (put a backslash in front of the space) but either I did not understand it properly or it does not work in that context. What I need is to specify a mount with a line like
"/mnt/HD_a2/Data/Time_Capsule" but to a folder called "Time Capsule". Yes I could change the name of the folder but then I would have to set up a whole heap of other connections again. Any suggestions?
3) The only other thing is that somehow large parts of the ffp directory has become set as no access rather than read only as it was previously. I am by no means sure that this is connected with setting up AFP as I have also been sorting another unrelated problem with Transmission. While not a problem per se it is inconvienient. I understand that I can change permissions of a directory with "chmod ugo+rx directory-name" run this as root or sudo BUT I also realise that I can very easily screw things up completely with this.
PLEASE tell me if I am reading this wrongly from the linux tutorial I consulted.
Thanks again for the help.
LorenzoG
Offline