Unfortunately no one can be told what fun_plug is - you have to see it for yourself.
You are not logged in.
All,
Thanks for all the work into getting the utilities compiled and working. I've successfully been running afpd on my dns-323, and it has resolved some file locking issues that drove me nuts with smbd and snow leopard / lion. Particularly the "file in use" errors while copying and deleting.
There have been a few weird issues, that I'm having trouble resolving.
1) HDD0 does not hibernate now that it is shared with afpd. I'm running ffp from a usb drive mounted read-only, and the drives have always pretty much stayed in hibernate. With afpd (and cidb) running, the drives will stay down so long as nothing is connected. However, with my MBP connected to the volume the HDD appears to pretty much wake up every 30 min or so. afpd log file is on the ramdisk, so it shouldn't be that.
2) Twice, i've noticed "TCP: Drop open request port : ####" in dmesg from a windows client. Now you may say, why would afpd and a windows client cause this? Good question, but I've never seen this before afpd was running.
One more question. Netatalk's website shows that dbus is required for running. I know support can be turned on/off while compiling, but what features would dbus enable?
Thanks.
Offline
rx2 wrote:
I am running netatalk 2.2.0 and avahi 0.6.30 - but as other have mentioned performance is pretty slow in 2 regards. 1st is when Lion is connecting to the AFP share - takes 1-2 minutes. And second is Time Machine is extremely slow. Its upto 3.9Gigs after 3.5hours and still going.........
Does anyone have any suggestions on speeding up the initial connection, or can comment on how long there initial backups are taking??
I have figured out a why to increase the speed of the connection a lot if you don't care about security:
/mnt/HD_a2/ffp/etc/netatalk # cat afpd.conf | grep tm-share
"tm-share" -tcp -uamlist uams_guest.so -setuplog "default log_info /ffp/var/run/afpd.log" -guestname nobody -cnidserver
The above enables the guest account of nobody
And them my volume share is:
/mnt/HD_a2/ffp/etc/netatalk # cat AppleVolumes.default | grep tm | grep -v "\#"
/mnt/HD_a2/tm tm options:tm,upriv,usedots
and my afpd.sh files is:
insmod /ffp/sbin/usb-storage.ko
sleep 10
/ffp/sbin/cnid_metad
sleep 10
mount /dev/sdc1 /mnt/HD_a2/tm
sleep 20
. /ffp/etc/ffp.subr
name="afpd"
command="/ffp/sbin/$name"
run_rc_command "$1"
Now, when I connect it immediately pops up the user box and I just choose 'guest'.
Hope this helps.
BTW, thanks you so much to everyones input because this is awesome once it is working. In my case, I have a 1TB drive hanging off the DNS-323 using usb-storage.ko and I am backing up to that drive.
Offline
Is anybody interested in compiling the new netatalk 2.2.2 for the rest of the community? I don't have a build environment :-).
Offline
I recently tried to upgrade to netatalk 2.2.1 as posted in this thread. Got everything working but now I see in the /var/log/messages file the following warning message: "volume "TimeMachine" does not support Extended Attributes, using ea:ad instead". What does this mean and is it a problem? I don't this warning message or issue existed before I upgraded. Previously I was using the version of netatalk that was installed from fonz' directory (2.0 I believe). Thanks!
Offline
I'm happy to see so many of you have succeeded at getting time machine backups to work from Lion to the fun-plugged DNS-323. Sadly, I'm not one of them. I followed rr-m's instructions from page 1 of this thread and was able to install the key packages (including the now updated versions):
avahi-0.6.30-1.tgz
db51-5.1.25-1.tgz
libdaemon-0.14-1.tgz
libgcrypt-1.5.0-1.tgz
libgpg-error-1.10-1.tgz
netatalk-2.2.0-1.tgz
I made changes to the files in /ffp/etc:
AppleVolumes.default
afpd.conf
The good news is I'm able to mount the volumes without error. I can create, copy and move files between my lion system and dns-323.
I created a fresh, new sparsebundle following the instructions at: http://www.horto.ca/?p=12 and moved it the DNS-323
Backups, however, aren't working. After the first step of "looking for backup disk..." I'm getting the message that others have reported: "The network backup disk does not support the required AFP features."
What might the "required AFP features" be? Can y'all suggest where I should focus my troubleshooting?
Thanks
BILL
Offline
I'm having the exact same problem as pidginbil with my DND-323. Installed the key packages. Made changes to the files. I can browse to the DNS-323 using command-k and i can now see it as an option to select in Time Machine preferences but get that same error: The network backup disk does not support the required AFP features
Console shows the volume does not support TM Lock Stealing. Backup failed with error: 45.
I haven't created a sparsebundle, I didn't think you had to do that now? I feel I'm pretty close to getting this working. I'd love to save $100 by not having to go out and buy a DNS-320!
UPDATE: looks like it is working!!! the documentation on what to edit our AppleVolumes.default is what the problem was. If anyone needs assistance I'd say to post the contents of theirs. Learned a bit about commenting and uncommenting though!
Last edited by cyberphox (2012-01-20 23:48:16)
Offline
Guys,
I updated avahi and netatalk and everything is working well, the 323 shows up in my shared devices in finder, however when I connect I see:
"Something wrong with the volume's CNID DB, using temporary CNID DB instead.Check server messages for details. Switching to read-only mode."
I can browse around just fine, but I'd like to get rid of that message and be able to make some modifications in there Any thoughts?
Offline
I was able to get it working now, I was never starting the CNID DB daemons and had trouble finding how you even did that. Before AFP starts i run the following:
/ffp/sbin/cnid_metad
/ffp/sbin/cnid_dbd
Once that was done there was still some really slow connections going on... The winners for me were:
afpd.conf:
"dlink-dns323" -uamlist uams_dhx.so,uams_clrtxt.so -setuplog "default log_info /ffp/var/run/afpd.log" -cnidserver
and AppleVolumes.default:
/mnt/HD_a2 Volume_1 options:usedots,upriv
All seems to be working really well now! Thanks to everyone in this thread! (I don't use time machine over this, so I can't comment on that, but it connects fast, and shows up in the shared devices which is really all i wanted!)
Offline
Not sure what happened to my other post....
Here's my working Time Machine setup on DNS-323 with Lion:
Install:
funpkg -i libgpg-error-1.10-1.tgz funpkg -i libgcrypt-1.5.0-1.tgz funpkg -i db51-5.1.25-1.tgz funpkg -i netatalk-2.2.1-1.tgz
/ffp/etc/netatalk/AppleVolumes.default:
/mnt/HD_a2/yourBackupVolume options:usedots,upriv,tm
/ffp/etc/netatalk/afpd.conf:
- -tcp -noddp -uamlist uams_dhx2.so,uams_clrtxt.so -setuplog "default log_info /ffp/var/run/afpd.log" -cnidserver localhost
I think you also have to make sure cnid_metad gets started before afpd.
Create /ffp/start/cnid_metad.sh and chmod +x it:
#!/ffp/bin/sh # PROVIDE: cnid_metad # REQUIRE: LOGIN . /ffp/etc/ffp.subr name="cnid_metad" command="/ffp/sbin/$name" run_rc_command "$1"
Also in /ffp/start/afpd.conf change this line:
# REQUIRE: LOGIN
to:
# REQUIRE: cnid_metad
If you want avahi magic, so that your backup volume shows up in the Time Machine as if it were an Apple Time Capsule:
Install:
funpkg -i avahi-0.6.30-1.tgz funpkg -i libdaemon-0.14-1.tgz
Create /ffp/start/avahi.sh and chmod +x it:
#!/ffp/bin/sh # PROVIDE: avahi # REQUIRE: SERVERS . /ffp/etc/ffp.subr name="avahi" command="/ffp/sbin/avahi-daemon" avahi_daemon_flags="-D -s" required_files="/ffp/etc/avahi/avahi-daemon.conf /ffp/etc/avahi/hosts" start_cmd="avahi_start" avahi_start() { # need avahi user and group for priviledge separation if ! grep '^avahi:' /etc/passwd >/dev/null; then echo 'avahi:x:50:50:Avahi Daemon:/no/where:/bin/false' >>/etc/passwd fi if ! grep '^avahi:' /etc/shadow >/dev/null; then echo 'avahi:*:14493:0:99999:7:::' >>/etc/shadow fi if ! grep '^avahi:' /etc/group >/dev/null; then echo 'avahi::50:avahi' >>/etc/group fi proc_start $command } run_rc_command "$1"
Create /ffp/etc/avahi/services/afpd.service:
<?xml version="1.0" standalone="no"?><!--*-nxml-*--> <!DOCTYPE service-group SYSTEM "avahi-service.dtd"> <service-group> <name replace-wildcards="yes">%h</name> <service> <type>_afpovertcp._tcp</type> <port>548</port> </service> </service-group>
Create /ffp/etc/avahi/services/adisk.service:
<?xml version="1.0" standalone="no"?><!--*-nxml-*--> <!DOCTYPE service-group SYSTEM "avahi-service.dtd"> <service-group> <name replace-wildcards="yes">%h</name> <service> <type>_adisk._tcp</type> <port>9</port> <txt-record>dk0=adVN=yourBackupVolumeName,adVF=0x81</txt-record> </service> </service-group>
With all of that in place, Lion will happily back up to the DNS-323.
If you run into trouble, check /ffp/var/run/afpd.log
Offline
Thanks, sneaky, for the additional information, I imagine it will be helpful to several people
Offline
Noticed that this is DNS-323/321 discussion, but gets so close to my question.
So got my hands on DNS-313 about week ago. Mainly my goal is get it to work with OS X Lion as a network share/drive. So right now don't need to get Time Machine function, just plain access to it (read/write). I don't have any data on it, so I can mess around with it.
I have been reading s**t load of forums and different workarounds how to get this going, but still no luck.
Right now think I've been messing around with this so much that now I'm totally lost and confused...
So far I have successfully installed ffp 0.7 oarm to it and installed required packages to it.
My concern is that I have not made right modifications to AppleVolumes.default and afpd.conf files.
Right now my afpd.conf file is in its default. Cause I have understood that default - -tcp -noddp -uamlist uams_dhx.so,uams_dhx2.so -nosavepassword would be enough. Or is it?
Also replaced AppleVolumes.default file default ending ~/ with /mnt/HD_a2/Volume_1. Shold this be enough or is this where I go wrong?
And by the way (maybe a totally stupid question), should I format Volume_1 from NTFS to ext2, ext3 etc. to get things go better?
If anyone can show me guidelines with this one, would appreciate it.
Offline
Hi, I have a DNS-323 and just got a Mac with Lion and wanted to setup Time machine. I followed instructions in http://www.noulakaz.net/weblog/2009/11/ … e-dns-323/ and then re-did it, only using the 0.7 area they mentioned (http://www.inreto.de/dns323/fun-plug/0.7/). I also did the items mentioned by sneakybits in post 84.
I'm using netatalk 2.2.2.
I mount the volume with afp://#.#.#.#/Volume_1 and, when asked I log in as nobody, with the password that I created per the instructions in the other link. It mounts the volume, but as read only, with the message about something wrong with the CNID DB, and it used a temp CNID DB instead. I also see the volume in TM.
In the afpd.log, I see this:
Feb 15 18:58:58.425922 afpd[1763] {uams_dhx2_passwd.c:270} (I:UAMS): DHX2 login: nobody
Feb 15 18:59:37.661369 afpd[1763] {auth.c:270} (N:AFPDaemon): AFP3.3 Login by nobody
Feb 15 18:59:37.730949 afpd[1763] {volume.c:2221} (W:AFPDaemon): volume "Volume_1" does not support Extended Attrib
utes, using ea:ad instead
Feb 15 18:59:37.731730 afpd[1763] {volume.c:2221} (W:AFPDaemon): volume "Volume_2" does not support Extended Attrib
utes, using ea:ad instead
Feb 15 18:59:37.733729 afpd[1763] {volume.c:2128} (I:AFPDaemon): Volume /mnt/HD_a2 use CNID scheme dbd.
Feb 15 18:59:37.734063 afpd[1763] {volume.c:2133} (I:AFPDaemon): CNID server: localhost:4700
Feb 15 18:59:37.896317 afpd[1771] {dsi_tcp.c:212} (I:DSI): AFP/TCP session from #.#.#.#:49940
Feb 15 18:59:38.026118 afpd[1763] {socket.c:157} (E:AFPDaemon): read: Connection reset by peer
Feb 15 18:59:38.051659 afpd[1763] {socket.c:157} (E:AFPDaemon): read: Connection reset by peer
Feb 15 18:59:39.072828 afpd[1763] {socket.c:157} (E:AFPDaemon): read: Connection reset by peer
Feb 15 18:59:40.148840 afpd[1763] {socket.c:157} (E:AFPDaemon): read: Connection reset by peer
Feb 15 18:59:59.198662 afpd[1763] {cnid_dbd.c:425} (E:CNID): transmit: Request to dbd daemon (db_dir /mnt/HD_a2) timed out.
Q1: Is it correct to log in as nobody, so should I setup an account that matches my user account on the Mac?
Q2: Should I set this up to mount Volume_1, or should I change it to mount some directory under that Volume? I want to use the DNS for PC and Mac backups, so I have folders on the Volume.
Q3: Any pointers as to what I may have done wrong, in my setup?
Thanks in advance!
Offline
I was reading some posts and I tried deleting the .AppleDB file and then attempted to reconnect and it did w/o complaining about CNID error. I'm trying Timemachine and it sees the volume. I picked it for the backup location, and it asked for my password, and then user/password on the server. Getting a busy spinning circle thingy for quite a while (minute or so), but finally completed. It's starting a backup now. Fingers crossed.
Questions:
1) I tried dbd -s at one point, and I get a Seg fault. Any idea what is up there?
2) I think I was reading that it would suck up all the space on the volume. Is that true (I have other folders used for backups of other machines)?
3) Is there a way to limit the space it uses? If so, how?
4) I ended up using this for afpd.conf, is it OK?
- -tcp -noddp -uamlist uams_randnum.so,uams_dhx2.so -nosavepassword -setuplog "default log_info /ffp/var/run/afpd.
log" -cnidserver localhost
Thanks in advance!
Offline
Well, backup failed. I got the message about a problem with the CNID DB and using temp one in read-only mode. Then, another dialog saying the backup image /Volumes/Time_Capsule-1/Seraph.sparsebundle could not be created, error 30.
Any ideas?
Offline
pcm wrote:
Well, backup failed. I got the message about a problem with the CNID DB and using temp one in read-only mode. Then, another dialog saying the backup image /Volumes/Time_Capsule-1/Seraph.sparsebundle could not be created, error 30.
Any ideas?
I had the same issue with my dns-323 and ftp 0.7. Recompiled db version solves the problem: more info. Download link: db5-5.2.36-oarm-1.txz
Offline