normtown Posted August 19, 2013 Share Posted August 19, 2013 I'm noticing that I can connect to my server by its hostname ("tower.local") when the server is started, but after "a while" (usually no longer than 30 minutes) the hostname is no longer resolvable. I can still connect by IP. Has anyone else seen this behaviour? I'd like to continue to used DHCP for IP assignment and not resort to hacking my hosts file. This should be something that works without resorting to those kind shenanigans This is my first unRAID build. I'm running 5.0 rc-16. Link to comment
Bitz69 Posted August 19, 2013 Share Posted August 19, 2013 I'm noticing that I can connect to my server by its hostname ("tower.local") when the server is started, but after "a while" (usually no longer than 30 minutes) the hostname is no longer resolvable. I can still connect by IP. Has anyone else seen this behaviour? I'd like to continue to used DHCP for IP assignment and not resort to hacking my hosts file. This should be something that works without resorting to those kind shenanigans This is my first unRAID build. I'm running 5.0 rc-16. Pretty typical here.. happens daily. I've just accepted it. Link to comment
trurl Posted August 19, 2013 Share Posted August 19, 2013 Have you tried Settings - SMB - Local master - Yes? After setting this you may have to wait a while, or just try rebooting all the other machines on the network. Link to comment
Bitz69 Posted August 20, 2013 Share Posted August 20, 2013 Have you tried Settings - SMB - Local master - Yes? After setting this you may have to wait a while, or just try rebooting all the other machines on the network. yes, been through all of that, it's just never reliable. it'll work sometimes but usually stops working on it's own. Link to comment
EGOvoruhk Posted August 21, 2013 Share Posted August 21, 2013 Pretty typical here.. happens daily. I've just accepted it. Is this pretty much the answer so far? To just accept it for now? I've been looking around to see if there's a solution, but most threads kind of dead end with nothing, or someone saying to edit the host file. I haven't seen anyone solve it I thought I was going mad when it all of a sudden stopped working. I'm (oddly) glad to see that I'm not the only one. I do hope it's looked into though Link to comment
normtown Posted August 22, 2013 Author Share Posted August 22, 2013 My guess is that there's something funny going on with the NetBIOS integration. I'm guessing that because I'm just running a simple LAN with a few machines connected to a router. For my unRAID server to disappear after a little while it would have to stop broadcasting its name to the other hosts on the LAN. Was this happening in 4.x? Have there been any changes in this area? Link to comment
unevent Posted August 22, 2013 Share Posted August 22, 2013 I've been getting this since 16c with avahi loaded. I don't use AFP or anything apple, yet avahi is running and creating tower-smb and tower.local crap. Link to comment
normtown Posted August 22, 2013 Author Share Posted August 22, 2013 I do use AFP. Creating the Tower and Tower-SMB crap is exactly what I want it to do This is affecting more than just the creation of Tower and Tower-SMB. As I said before, it seems as though the unRAID server is not broadcasting its NetBIOS hostname to the LAN. I can't even ping the server with its hostname. The only thing that seems to make it broadcast again is restarting my router. I'm guessing that the process of getting a new IP address from my router's DHCP server causes the unRAID server to broadcast its hostname. I'm guessing there's a TTL on that? Link to comment
unevent Posted August 22, 2013 Share Posted August 22, 2013 I started experiencing the same problem you describe in both my Windows and Linux system over SMB when avahi was included in 16c. Those that don't have AFP enabled should not suffer from having avahi enabled by default. Link to comment
dgaschk Posted August 22, 2013 Share Posted August 22, 2013 tower.local is an ahavi name. SMB clients should connect to the shares with \\tower and to the management GUI with //tower Link to comment
unevent Posted August 22, 2013 Share Posted August 22, 2013 tower.local is an ahavi name. SMB clients should connect to the shares with \\tower and to the management GUI with //tower Yes, I know. My point was I have been experiencing th same problems as the OP (host name not working at times), but also included the comment that since I do not use AFP I should not see 'Tower-SMB' and 'Tower.local' in my network views because I the wife and kids that access unRAID from computers and having extra stuff to click tot he Tower does not go over well. If it is for AFP only then if AFP is not enabled it should not be there. Link to comment
normtown Posted August 23, 2013 Author Share Posted August 23, 2013 I understand your frustration regarding the advertised services from avahi. However, that is not really what this thread is about. If you have any solutions or clues about the root cause of the original issue I reported, please feel free to contribute. It sounds like the integration with avahi coincides with this hostname behavior. Are there any unRAID devs that can speak about the integration? Link to comment
EGOvoruhk Posted August 23, 2013 Share Posted August 23, 2013 It sounds like the integration with avahi coincides with this hostname behavior. I just tried running a fresh install of 5.0-rc15a, which according to the changelogs is the release before AVAHI runs regardless of AFP status, but it's still getting the hostname issues Link to comment
dgaschk Posted August 23, 2013 Share Posted August 23, 2013 Avahi is not the issue. SMB clients are not effected by Avahi. Link to comment
sureguy Posted August 24, 2013 Share Posted August 24, 2013 Can you try booting from a bootable Linux CD/DVD/USB and then try again? I can ping my server name from all my workstations (win xp, win 8, ubuntu, etc) without issue. This would either eliminate or implicate Windows as the cause of the problem. I'm using SimpleFeatures with the following SMB settings: Enable SMB: Yes(workgroup) Workgroup: workgroup Local Master: Yes Link to comment
dgaschk Posted August 24, 2013 Share Posted August 24, 2013 I do use AFP. Creating the Tower and Tower-SMB crap is exactly what I want it to do This is affecting more than just the creation of Tower and Tower-SMB. As I said before, it seems as though the unRAID server is not broadcasting its NetBIOS hostname to the LAN. I can't even ping the server with its hostname. The only thing that seems to make it broadcast again is restarting my router. I'm guessing that the process of getting a new IP address from my router's DHCP server causes the unRAID server to broadcast its hostname. I'm guessing there's a TTL on that? Sounds like a router issue. Link to comment
EGOvoruhk Posted August 24, 2013 Share Posted August 24, 2013 This would either eliminate or implicate Windows as the cause of the problem. It's not Windows. I can't ping from any device. Windows, OS X, Android, etc. It's also not the UnRAID hardware, since I could ping the Server 2012 install that was on it days before I did a little more testing today, and I think it's a problem with whatever scripts run when setting a static IP in the web GUI. I did another fresh install of rc16c, and after boot, I can ping Tower just fine. I did nothing but change the IP from DHCP to Static, reboot, and as soon as it comes up, I can't ping Tower anymore. I also don't see the hostname in the device list on my router (running Tomato). So I change the IP back to DHCP, reboot, and as soon as it finishes booting, the device pops up in Tomato's device list, and I can ping Tower again Just for a little confidence, I used a 2nd flash drive, installed UnRAID rc16c on it, and then plugged it into an older Thinkpad I've got. I then performed the exact same steps as above, and got the same results. Upon initial boot, I could see and ping Tower. I then changed the IP to Static (a different IP than the 1st UnRAID machine), rebooted, and when it came back up, it was again no longer in Tomato's device list and no longer pingable. Changed IP back to DHCP, rebooted, and it's back in the device list and pingable again normtown, no doubt if you performed the same steps, you would see the same results I'm seeing Since I need a static IP, and I also want to use unMENU (which only works by browsing via the hostname), the best course of action is to use IP reservations in my router, which isn't the worst thing in the world, but this is still a problem that needs addressing Link to comment
normtown Posted August 25, 2013 Author Share Posted August 25, 2013 I can't ping tower from any host in my network either. @EGOvorhuk I'm a little confused by your test. My networking knowledge is admittedly limited. If you set a static IP in the web GUI but don't reserve that IP in your DHCP server, should your router know about it? It's not clear to me that the behavior you're seeing is unexpected. Can anyone educate me on this? Also, when I reboot (using DHCP for IP assignment) I see Tower in my device list and it is pingable via tower.local at first. After a little while (not sure how long exactly--5 to 30 minutes, maybe?) tower.local becomes unpingable. Link to comment
normtown Posted August 25, 2013 Author Share Posted August 25, 2013 I just verified that I get the same behavior when using a static IP address. Tower appears at first and is pingable as tower.local, then it disappears after a little while and is no longer pingable by the hostname. Link to comment
dgaschk Posted August 25, 2013 Share Posted August 25, 2013 Attach a syslog that includes a disconnect period. Link to comment
normtown Posted August 27, 2013 Author Share Posted August 27, 2013 Attached! syslog.txt Link to comment
dgaschk Posted August 27, 2013 Share Posted August 27, 2013 What does "ethtool eth0" and "ifconfig" show? Link to comment
unevent Posted August 27, 2013 Share Posted August 27, 2013 Attached! Problem looks to be in AFP configuration and how it handles clients going to sleep. It is something Tom will most likely need to look into and is probably already on his list for after 5 is released. Something to chew on in the meantime http://osdir.com/ml/network.netatalk.devel/2003-03/msg00023.html Link to comment
drawz Posted August 28, 2013 Share Posted August 28, 2013 What happens if you go to a command line (on your PC, not the unRAID box) and type in "nslookup tower" (replace tower with the name of your unraid machine). If it can't find the IP address, this could be a problem with your routers handling of hostnames. You can try this right when your box boots and the webgui works and then try again when it goes down. What router are you using? Link to comment
normtown Posted August 29, 2013 Author Share Posted August 29, 2013 Firstly, on the unRAID server, I ran ethtool and ifconfig: ~# ethtool eth0 Settings for eth0: Supported ports: [ TP ] Supported link modes: 10baseT/Half 10baseT/Full 100baseT/Half 100baseT/Full 1000baseT/Full Supports auto-negotiation: Yes Advertised link modes: 10baseT/Half 10baseT/Full 100baseT/Half 100baseT/Full 1000baseT/Full Advertised pause frame use: Symmetric Receive-only Advertised auto-negotiation: Yes Speed: 1000Mb/s Duplex: Full Port: Twisted Pair PHYAD: 0 Transceiver: internal Auto-negotiation: on MDI-X: Unknown Supports Wake-on: pumbg Wake-on: g Current message level: 0x00000033 (51) Link detected: yes ~# ifconfig eth0 Link encap:Ethernet HWaddr bc:5f:f4:83:ef:71 inet addr:192.168.0.100 Bcast:192.168.0.255 Mask:255.255.255.0 UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:2467498 errors:0 dropped:198 overruns:0 frame:0 TX packets:923886 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:3280671312 (3.0 GiB) TX bytes:64237367 (61.2 MiB) Interrupt:46 Base address:0x2000 lo Link encap:Local Loopback inet addr:127.0.0.1 Mask:255.0.0.0 UP LOOPBACK RUNNING MTU:65536 Metric:1 RX packets:917 errors:0 dropped:0 overruns:0 frame:0 TX packets:917 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:0 RX bytes:76862 (75.0 KiB) TX bytes:76862 (75.0 KiB) Link to comment
Recommended Posts
Archived
This topic is now archived and is closed to further replies.