SJ_UnderWater

Timemachine On Lion With Netatalk

353 posts in this topic

Because of the myriad formatting issues, and the problems they are causing users this solution has been moved to Tutorials. Please download the files there, making sure to only unzip them, they were zipped to get around the file restrictions here.

(WHSv1)

Share this post


Link to post
Share on other sites

Upgrade to a WGS Supporter Account to remove this ad.

Looks very promising.....possible enough for me to hold out on trying to get linux running on my EX490....

Share this post


Link to post
Share on other sites

demo, if you could do some performance testing for me, that would be great. there are some experimental options that can increase the speed slightly, but at the cost of reliability (using "cnidscheme:cdb" in AppleVolumes.default, for instance). Since nearly everyone will be using this solution exclusively for backups I chose the most robust configuration. On my small Atom machine i clocked about 2MiB/s, I'm sure you'll have a better go.

Share this post


Link to post
Share on other sites

coLinux doesn't run on 64bit windows yet...so we should note that this is for 32bit windows machines only at this stage.

Share this post


Link to post
Share on other sites

im working on getting ArchLinux working on QEMU for Windows, which is 32/64bit, but should only be used on machines that support hardware virtualization (Intel's VT-x). There are so few virtualization options that I'm trying to make sure they're efficient before I recommend them. coLinux is definitely the fastest I've seen.

--edit

i forgot to mention that most other vm solutions require the TAP option with bridging, meaning network bridging has to be manually reinstalled on the WHS (it's uninstalled by default), they dont use winpcap like colinux

Share this post


Link to post
Share on other sites

i don't think it's possible because you're basically setting up a second operating system. I don't think it's even possible to script coLinux itself. The only real possibility would be setting up a finished archlinux image that users could download, but the size would be pretty huge. I'll think about it when I finish QEMU for WHS 2011

Share this post


Link to post
Share on other sites

I can confirm that this does work....to a point...

The time machine does show in lion, no mods required, and can connect, however in my case, after writing about 200-400mb, it stops, and then gives a "disk unmounted" error.

From what I can tell, it's like the VM is hard locking....I can't login, can't type anything on the console...

Looks like its usb hdd timemachine for the time being for me......

Share this post


Link to post
Share on other sites

I can confirm that this does work....to a point...

The time machine does show in lion, no mods required, and can connect, however in my case, after writing about 200-400mb, it stops, and then gives a "disk unmounted" error.

From what I can tell, it's like the VM is hard locking....I can't login, can't type anything on the console...

Looks like its usb hdd timemachine for the time being for me......

Has anyone gotten this to work and complete a TM backup? Still running 10.6.8 although since I upgraded from 10.6.7 to 10.6.8, I've had nothing but problems with Time Machine. Hoping AFP would solve that. Was hoping to use AFP and then upgrade to Lion.

No experience with Linux, but I'm willing to adventure a bit (my Mac is really a Hackintosh) and the instructions look pretty clear, but don't want to invest the time if this really doesn't work reliably.

Share this post


Link to post
Share on other sites

demo and I are working out a few edge cases, but we haven't decided the final resting place of the backups folder. The original choice, an unshared folder in D:\shares, didn't turn out so well since that whole folder is indexed for search. I've put the question out to a few experts and they'll get back to me, but the actual file transfer and backup works fine. You don't need TMShowUnsupportedNetworkVolumes, or to manually create an image, and it works smoothly. I'm also making a few small changes to the script for simplicity/speed.

1 person likes this

Share this post


Link to post
Share on other sites

demo and I are working out a few edge cases, but we haven't decided the final resting place of the backups folder. The original choice, an unshared folder in D:\shares, didn't turn out so well since that whole folder is indexed for search. I've put the question out to a few experts and they'll get back to me, but the actual file transfer and backup works fine. You don't need TMShowUnsupportedNetworkVolumes, or to manually create an image, and it works smoothly. I'm also making a few small changes to the script for simplicity/speed.

Awesome. Looks like this might tax my HP EX-490 celeron cpu running a virtualized os layer. Good excuse to upgrade to E7500 or something + memory upgrade. Keep us posted on your good work.

Share this post


Link to post
Share on other sites

it probably will tax a celeron, but the RAM usage is only about 9MiB

Share this post


Link to post
Share on other sites

Hi everyone, I'm trying this for the second time, but always get stuck at step 8.. After the last command the script runs but says:

Generating locales....

en_US.UTF-8... done

Generation complete.

And just stays there.. Nothing more.. Not asking for password or anything else..

Any idea of what the problem could be?

thanks in advance!

Share this post


Link to post
Share on other sites

ive made a minor change to the script, but my first impression is that your vm doesn't have an internet connection. when you say it does nothing, do you mean it returns to the prompt? if it _does_ return to the prompt, it exited because of an error (like no internet connection). if it _doesn't_ return to the prompt, it's still working on something. I suggest you open the WHS's task manager (right-click the taskbar) and watch the CPU usage.

Share this post


Link to post
Share on other sites
Thanks for the fast reply! It goes back to the prompt.. However, I do have internet connection as I can surf the web using firefox on the WHS.

I think it will help to troubleshoot if I write down in bold what I've done in each step:

Login to the WHS. I used remote desktop connection from my macbook pro.Download coLinux, and install it to C:\coLinux (only FLTK, NDIS, and Root Filesystem are necessary), choosing ArchLinux when prompted. Download the coLinux on my mac, copied it to the WHS, installed it on C:\coLinux as instructed. Archlinux downloads just fine.Download and expand the 64MB swap file, and toporesize. Make sure the ArchLinux image (.fs) is named "arch.fs", the swap file is named "arch.swap", and both are in C:\coLinux. You may unzip toporesize to a subdirectory. Downloaded the swap file and toporesize on the mac. Extracted both using the unarchiver and named them according to instructions. Once extracted, I add the extension .swap to the swap file. Move both files to the WHS and to the coLinux folder. Copied the ArchLinux.7z from the coLinux folder to the mac in order to extract it using the unarchiver. Copied it back and named arch.fs. Save both scripts below (fixarch.sh, colinux.bat) and the netatalk package to C:\coLinux. Using notepad, pasted the scripts and saved them using the corresponding extensions. Downloaded the netatalk package and moved it as is to the coLinux folder.Run toporesize from "toporesize.bat", choose find file and navigate to arch.fs, and resize to 384MiB or greater (the finished distribution is ~278MiB but needs more in between). Done exactly this, resized to about 540.With all of the files in place, create a shortcut to colinux-daemon.exe, and change it to the shortcut below. Done this, no problem here.Run the shortcut you created. ArchLinux will start, but in an unfinished state. Archlinux runs, all fine.Run each of the following commands (separated by semicolons): root; mount /mnt/win; cat /mnt/win/coLinux/fixarch.sh | tr -d "\r" >/tmp/fixarch.sh; chmod u+x /tmp/fixarch.sh; /tmp/fixarch.sh Run every single command, but like I explained in the previous command, it returns to the prompt.

Thanks for the help!

Share this post


Link to post
Share on other sites

Just tried it with the new script.. It gives me an error saying "no internet connection!"

I can confirm I do have internet connection as like I said i'm connecting to the web using firefox.

However, could it be that the problem is with how I have configured my network adapters? I'm using a 680i motherboard with two network adapters that are teamed up (NIC teaming) in order to provide a 2gbps connection. The connection that i'm using is Local Area Connection 4.

Just to try, I changed the shortcut to reflect the Local Area Connection 4 but with no success..

Any recommendations?

Share this post


Link to post
Share on other sites

i don't know that NDIS supports bonding, but since you changed the shortcut to "Local Area Connection 4" (what are the other 3?) and it still didn't work, you could try deleting the arch.fs image and try again, but I don't know that that will fix anything, coLinux is prepared to support different configurations at each boot.

You could try the other semi-fast connection option WinPCAP (install first!) with the following code:

eth0=pcap-bridge,"Local Area Connection 4"

for reference the speeds are:

NDIS > WinPCAP > TUNTAP > SLIRP

Share this post


Link to post
Share on other sites

i would just like to point out to others that this is a very special case, very few people use link aggregation/bonding with a WHS, and the drivers/subsystems necessary don't work with all applications, like coLinux which needs to "piggyback" on a network connection. NDIS is the lowest-level piggyback available, uses Windows' own native subsystem, and has great performance, but expects a "normal" network connection.

WinPCAP is also pretty fast and may be more compatible with unusual configurations.

Share this post


Link to post
Share on other sites

Hi there, I am getting this error and any help would be appreciated:

Generating locales...

en_US.UTF-8... done

Generation complete.

:: Synchronizing package databases...

error: failed retrieving file 'core.db.tar.gz' from mirror.rit.edu : Not Found error: failed to synchronize core: Not Found

error: failed retrieving file 'extra.db.tar.gz' from mirror.rit.edu : Not Found error: failed to synchronize extra: Not Found

error: failed retrieving file 'community.db.tar.gz' from mirror.rit.edu : Not Found

error: failed to synchronize community: Not Found

error: failed to synchronize any databases

Thanks

Share this post


Link to post
Share on other sites

i can't say for sure what the exact problem is, just that you don't have access to ArchLinux's update server for some reason. Are you far away from rhode island, or have other connection issues?

Share this post


Link to post
Share on other sites

i can't say for sure what the exact problem is, just that you don't have access to ArchLinux's update server for some reason. Are you far away from rhode island, or have other connection issues?

thank you for a quick reply... I am in NYC and connection seems to be ok...

Share this post


Link to post
Share on other sites

Hi, I was just wondering if you'd consider this solution ready for mainstream use now.

Thanks!

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!


Register a new account

Sign in

Already have an account? Sign in here.


Sign In Now



Upgrade to a WGS Supporter Account to remove this ad.



  • Posts

    • I had no problem with Windows 10.  However, I just completed the anniversary update and now the "Connector" won't run to conclusion.  It backs itself out and reports an unspecified error.
    • Apologies if I'm being stupid, but I'm having great trouble understanding how Lightsout backup and the Server Essentials 2012 backup work together (or fight with each other). I find the Lightsout documentation on this very unclear, and various threads have left me no wiser. I'd appreciate some help! Here's the setup: Lightsout is running nicely on the server. Before backups were set up, I established that the server hibernated when the clients were switched off. The server woke up when a client was powered up. (This behaviour broke when backups were set up.) The server is Essentials 2012, with a number of client laptops: a mix of Windows 7, 8 & 10. The Essentials backup window is 22:00 - 07:00. No backup settings are currently made in Lightsout. Here's what I'd like to achieve: At the end of the working day all the clients are put to sleep or hibernate (or switched off completely, I don't care which). Ten minutes after that, Lightsout hibernates the server. Sometime during the night (22:00 - 07:00) the server wakes up, then wakes up the various clients in turn, backs them up, and puts them back to sleep / hibernate / shutdown (I don't care which). Here's what actually happens (when I look at the uptime chart the next morning): At the end of the day (about 17:15) the clients are put to sleep / hibernate. (I tell the team not to shut down their client laptops because I assume the server can't wake a completely shut down client - is that correct?) The server does not hibernate. Approximately 90 minutes after the clients were put to sleep / hibernate, they all wake up simultaneously (around 18:45, but I don't know why this happens). The clients then stay awake all night. Later, from 22:00, some of the backups take place (but not all of them). The backups are spread wide apart in the 22:00 - 07:00 window with lots of time when nothing is going on at all. The clients do not go back to sleep / hibernate after their backups. ...can anyone suggest what combination of settings I should be using to get the desired behaviour? Here are the sort of things that are confusing me, but I can't find definitive answers (and each experiment I do takes 24 hours to give any results!): I know I can specify clients' individual backup times in Lightsout, but I don't know how long a client's backup will take, so I could have backups overlapping each other and causing a traffic jam. Should I worry about this? If I set Lightsout backups, what do I set for the Essentails backup window? What if it's not long enough to get all the backups done? If I make settings in the Lightsout backup dialog do I have to complete everything in that dialog box? Does this make the Essentials 2012 backup settings redundant? What about the Essentials 22:00 - 07:00 backup window, does that matter any more? I tried setting client backup times in Lightsout with an end action of 'sleep'. The backups didn't happen at all and clients all spontaneously went to sleep in the middle of the working day (the server / clients clocks all match, I checked). Can Lightsout wake up a client laptop that is completely shut down? Thanks for reading this long post and sorry if these seem like stupid questions. I already have WHS2011 / Lightsout running perfectly on another setup (where my backups take place during the day), but nocturnal backups are a new level of complexity for me!
    • Christopher wrote:  << And to clarify, if you want multiple backup drives to be cycled offsite, you MUST connect all of them at the same time when you set up the backups. >> I'm not clear about this. To the best of my recollection, I never connected both of my backup drives to the server at the same time,  For test purposes, I just initiated a recovery (using the WHS Installation Disk) with each of them hooked up.  The one that has been offsite and the one that backed up last night were both found by the "Repair an existing installation" procedure.  Perhaps I'm misreading your clarification.
    • Please check if task scheduler is still working. If not use this tool here https://repairtasks.codeplex.com/ If that does not help you may try to reinstall WHS Connector.
    • My WHS v1 (home build) starting misbehaving and finally needed a new system disk.
      After lots of threads, trial and errors, AF and non-AF drives, and lack of the Server Reinstallation choice it has finally progressed.
      Weeks of frustration. My recent attempts proceeded well but all ended up without the rebuilding of the Storage Pool (6 drives).
      The drives are detected and show up as "not added" in Server Storage and as "non managed" in Disk Management.
      There is no C:\fs on the SYS volume.
      It appears that the RebuildPrimary stage of the reinstallation, mentioned by some posters, was not successful. Has anyone found a way to force the rebuilding of the Storage Pool on reinstallation?
      I have tried X: Files v1.5.5 and the original Microsoft Disk with the same outcome. [My next step will be to migrate the "shares" and "folders" manually - it would have be quicker to have done this initially.]
       
  • Popular Contributors