Hi folks,
I'm running CentOS 6.6 on my workstation. There I have a user meikel, and its home directory /home/meikel ist stored on a samba share on a server which also runs CentOS 6.6.
In the very beginning I added the mount options for that home directory into the /etc/fstab file which led to some problems, so I set up autofs for mounting the home directory.
The autofs seems to work for /home/meikel, but sometimes when I login at the GUI, I get error messages which look a little as if the home dir isn't available, I then also don't get my desktop background image. The two error dialogs which I get in that situation show the following messages:
"Could not update ICEauthority file /home/meikel/.ICEauthority"
and
"Es gibt ein Problem mit dem Konfigurationsserver. (/usr/libexec/gconf-sanity-check-2 beendet mit Status 256)"
Sorry for the german, it's not clear to me how to get it in english. It's something like "There's a problem with configuration server. (/usr/libexec/gconf-sanity-check-2 terminated with exit-code 256)"
To do some analysis I did the following:
- Restart workstation, wait until login screen appears
- use CTRL-ALT-F2 to go to a text console
- log in as root
- mount --> it seems that /home/meikel isn't mounted (and this is what I expect)
- ll /home/meikel/ --> it's fine, I can see alls the files and subdirectories
- ll /home/meikel/Dokumente/ --> it's also fine
- df -h /home/meikel --> now the mount is there
- mount | grep meikel --> yes, the mount is there
- use CTRL-ALT-F3 to go to another text console
- log in as user meikel --> I get an error saying "No directory /home/meikel!", but I'm logged in
- pwd --> it shows the root dir /
- id --> it shows uid=500(meikel) gid=500(meikel)
- ll -d /home/meikel --> drwx------ meikel:meikel
- echo ${HOME} --> shows /home/meikel
So the directory /home/meikel is definitely there. Can someone explain the error message "No directory /home/meikel!" and give any hints about possible problem causes? How can it say that the directory isn't here but I can see all the data?
Regards,
Meikel
Am 02.07.2015 um 20:16 schrieb Gordon Messmer:
No, I didn't set it. Now I've set it and when logging into text console I no longer have the problem with the home directory. That's fine now. Thanks a lot!
But the other problem remains, that very soon after startup I can't log into the GUI. I think the remaining problem doesn't has to do something with Samba/Autofs. It looks as if the network becomes ready very late (after autofs started), which perhaps would explain my problem.
In the /var/log/boot.log I see:
... Schnittstelle eth0 hochfahren: IP-Informationen für eth0 werden bestimmt ... fehlgeschlagen; keine Verbindung vorhanden. Kabel prüfen? FEHLGESCHLAGEN ... Netzwerkparameter einstellen ... OK NetworkManager-Daemon starten: OK Loading autofs4: OK automount starten: OK
In /var/log/messages I see:
Jul 5 16:36:08 meikel-pc kernel: ADDRCONF(NETDEV_UP): eth0: link is not ready Jul 5 16:36:23 meikel-pc kernel: ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
It takes 15 seconds between the two messages until it becomes ready. I have no idea why it first says that the link is not ready.
Any ideas?
Regards,
Meikel
On 07/05/2015 07:57 AM, Meikel wrote:
It's probably autonegotiation of link speed. I'm not sure why it'd take that long. I'd think the most likely explanation would be a bad cable. Could also be a flaky port on the switch, or a flaky Ethernet card. What brand is the local interface?
Am 05.07.2015 um 20:52 schrieb Gordon Messmer:
I'll check the cable (not clear how I can do that, probably using another cable and connecting to another outlet).
I think the local interface is directly integrated into the Intel Desktop Board DH87RL.
dmesg | grep eth e1000e 0000:00:19.0: eth0: registered PHC clock e1000e 0000:00:19.0: eth0: (PCI Express:2.5GT/s:Width x1) 00:22:4d:b0:e0:e1 e1000e 0000:00:19.0: eth0: Intel(R) PRO/1000 Network Connection e1000e 0000:00:19.0: eth0: MAC: 11, PHY: 12, PBA No: FFFFFF-0FF ADDRCONF(NETDEV_UP): eth0: link is not ready e1000e: eth0 NIC Link is Up 100 Mbps Full Duplex, Flow Control: None e1000e 0000:00:19.0: eth0: 10/100 speed: disabling TSO ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready eth0: no IPv6 routers present
lspci | grep Ether 00:19.0 Ethernet controller: Intel Corporation Ethernet Connection I217-V (rev 04)
It says that it has a speed of 10/100, but I definitely expect 1000 Mbps!
On 07/06/2015 10:00 PM, Meikel wrote:
I'll check the cable (not clear how I can do that, probably using another cable and connecting to another outlet).
Use a different cable to the same switch port, first. If you change both the cable and the switch port, you won't know if the problem was the cable or the switch port.
If the problem persists with a different cable, *then* try a different switch port.
e1000e 0000:00:19.0: eth0: Intel(R) PRO/1000 Network Connection e1000e: eth0 NIC Link is Up 100 Mbps Full Duplex, Flow Control: None
...
It says that it has a speed of 10/100, but I definitely expect 1000 Mbps!
Yeah, the low link speed is no surprise given the long negotiation time.
Google-fu also turns up a suggestion that the switch doesn't negotiate properly in some cases, and that the problem can be resolved by disabling power saving mode on the switch, in its management interface.
Am 08.07.2015 um 08:38 schrieb Meikel:
I think the next step for me to do is to check the ethernet cable(s).
Yes, it was a problem with the cable between the patch field and the outlet, I temporarily moved to another outlet and it works. And i double checked the defect cable with other devices and can see the same problems.
I still get the message
Jul 21 08:06:04 meikel-pc kernel: ADDRCONF(NETDEV_UP): eth0: link is not ready
but in the same second I then can see
Jul 21 08:06:04 meikel-pc kernel: e1000e: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: None Jul 21 08:06:04 meikel-pc kernel: ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
and as I can see the link speed is 1000 Mbps now. The other problem "Could not update ICEauthority file /home/meikel/.ICEauthority" has gone, so probably due to cabling problem the network became ready much to late and autofs wasn't able to mount /home when logging in. So now everything is fine.
Regards,
Meikel
What is the upstream switch? If it is a Cisco switch does the configuration have `spanning-tree portfast` enabled? http://www.cisco.com/c/en/us/support/docs/switches/catalyst-6500-series-swit...
On Jul 7, 2015, at 12:42 AM, Meikel meikel@fn.de wrote:
It's a D-Link DGS-1210-16 Switch.
http://www.dlink.com/-/media/Business_Products/DGS/DGS%201210%2048/Manual/DG...
What are the current settings on the switch for: Loopback Detection page 35 STP Global Settings page 38 STP Port Settings page 39
Under STP Global the "forward delay" defaults to 15 seconds... The same delay you're seeing.
-
Gordon Messmer -
Meikel -
Steven Tardy