Re: Update killed samba -
On 01/02/15 06:43, Tom H wrote: On Thu, Jan 1, 2015 at 4:01 PM, Bob Goodwin bobgood...@wildblue.net wrote: I ran a system update on an SL-7 server and that seems to have messed up samba. I can ssh into the system but can't mount it as a samba share. [root@box10 bobg]# mount //192.168.1.48/myshare /mnt/box48/ Password for bobg@//192.168.1.48/myshare: Unable to find suitable address. This has worked without a problem until now. Any thoughts on what might be my problem appreciated. Which packages were updated? Perhaps 100 items ending with: Dec 12 04:01:19 Installed: samba-4.1.1-37.el7_0.x86_64 Dec 12 04:02:25 Installed: alacarte-3.7.90-8.el7.noarch Dec 12 12:13:50 Installed: system-config-users-docs-1.0.9-6.el7.noarch Dec 12 12:13:51 Installed: system-config-users-1.3.5-2.el7.noarch Which makes Samba suspect. I also tried booting to the previous kernel. That made no difference. Did you change your samba config? No and it appears not to have been changed: [root@box48 bobg]# ll /etc/samba/smb.conf* -t -rw-r--r--. 1 root root 12175 Dec 22 12:16 /etc/samba/smb.conf -rw-r--r--. 1 root root 12130 Dec 22 11:48 /etc/samba/smb.conf~ -rw-r--r--. 1 root root 11702 Dec 15 10:24 /etc/samba/smb.conf.works -rw-r--r--. 1 root root 11630 Dec 12 09:03 /etc/samba/smb.conf.default It looks like it still has the entries I made last time. Did you change your iptables config? No, and stopping firewalld had no effect, I also tried setenforce 0. Did you reboot after the updates? Yes, a number of times now. Do the client or server logs give you more info? The server: [root@box48 bobg]# journalctl --since -60m .. snip .. Jan 02 10:01:01 box48 CROND[22259]: (root) CMD (run-parts /etc/cron.hourly) Jan 02 10:01:01 box48 run-parts(/etc/cron.hourly)[22262]: starting 0anacron Jan 02 10:01:01 box48 run-parts(/etc/cron.hourly)[22269]: finished 0anacron Jan 02 10:01:01 box48 run-parts(/etc/cron.hourly)[22271]: starting 0yum-hourly.cron Jan 02 10:10:01 box48 systemd[1]: Starting Session 210 of user root. Jan 02 10:10:01 box48 systemd[1]: Started Session 210 of user root. Jan 02 10:10:01 box48 CROND[22370]: (root) CMD (/usr/lib64/sa/sa1 1 1) Jan 02 10:20:01 box48 systemd[1]: Starting Session 211 of user root. Jan 02 10:20:01 box48 systemd[1]: Started Session 211 of user root. Jan 02 10:20:01 box48 CROND[22494]: (root) CMD (/usr/lib64/sa/sa1 1 1) While the client box shows: Jan 02 10:18:12 box10 kernel: CIFS VFS: Error connecting to socket. Aborting operation. Jan 02 10:18:12 box10 kernel: CIFS VFS: cifs_mount failed w/return code = -113 I'm curious about the difference in time. I normally block WAN in the router for the servers, I also have a separate NFS box. Does smbclient -L ... list the share(s)? [root@box48 bobg]# smbclient -L myshare/ Enter bobg's password: Connection to myshare failed (Error NT_STATUS_UNSUCCESSFUL) I even tried resetting the smb password. Thanks to all for responding, Bob -- http://www.qrz.com/db/W2BOD box10 Fedora-21/64bit Linux/XFCE
Re: Update killed samba -
On 2 January 2015 at 13:12, Bob Goodwin bobgood...@wildblue.net wrote: On 01/02/15 14:08, Stephen John Smoogen wrote: The ntp server needs to be able to talk to some server on the network to get time. The Fedora boxes might just be 'lucky' in keeping time if they have no access either. ntpq -p on each of the hosts will see how they are keeping time. Well it looks like SL-7 does not use ntp, it uses chrony instead and it apparently was not running, at least I had to systemctl restart chronyd|and then |systemctl enable chronyd to start it running. Then it began to show signs of life. Before that all the values reported were zero! [root@box48 bobg]# chronyc sources 210 Number of sources = 4 MS Name/IP address Stratum Poll Reach LastRx Last sample === ^+ clock.xmission.com1 61738 -522us[+2286us] +/- 312ms ^+ 44.s.dedikuoti.lt 2 61738-20ms[ -17ms] +/- 403ms ^+ bindcat.fhsu.edu 2 61738 +2636us[+5444us] +/- 369ms ^* lithium.constant.com 2 61738 +4799us[+7607us] +/- 381ms Only the servers are blocked from the internet since I figured they didn't require a connection but now the question of time synchronization arises. In fact my internet connection is derived from a satellite connection which has a system delay on the order of 800 ms. I would think that would offset me from the rest of the world by nearly one second if it matters ... Get yourself a little GPS gadget to lock into the satellite clocks on one of the boxes and call that server stratum 1. Aim the other boxes at Just found this.. which may help in other areas: http://www.linuxhomenetworking.com/wiki/index.php/Quick_HOWTO_:_Ch24_:_The_NTP_Server#.VKcREXsRRKo Interesting thought, Bob || -- http://www.qrz.com/db/W2BOD box10 Fedora-21/64bit Linux/XFCE -- Stephen J Smoogen.
Re: Update killed samba -
On 2 January 2015 at 10:15, Bob Goodwin bobgood...@wildblue.net wrote: On 01/02/15 11:57, Akemi Yagi wrote: On Fri, Jan 2, 2015 at 7:33 AM, Bob Goodwin bobgood...@wildblue.net wrote: Jan 02 10:18:12 box10 kernel: CIFS VFS: cifs_mount failed w/return code = -113 According to /usr/include/asm-generic/errno.h, return code -113 means /* No route to host */. Akemi No route to host for Samba I guess, I am accessing the server via ssh and that's how I gathered the data shown? As I said originally, I can't mount the samba share myshare from either Fedora-21 client. That began immediately after the system update, yum I presume? A notice popped up that there were updates that needed to be done and I simply told it to go ahead after I enabled it's WAN access. I don't think there's any difference between the yum and dnf updates any longer. When they first began testing dnf there were some differences, I no longer pay any attention to it, just do yum update with F-21 and assume the same for SL-7? Without WAN access I wonder how ntp keeps time, I assume it must get it from wherever it can find it on the LAN? The Fedora boxes have access and always have the correct time. Does the problem go away with yum downgrade samba That should move you back to an alternate RPM of samba. The ntp server needs to be able to talk to some server on the network to get time. The Fedora boxes might just be 'lucky' in keeping time if they have no access either. ntpq -p on each of the hosts will see how they are keeping time. Bob -- http://www.qrz.com/db/W2BOD box10 Fedora-21/64bit Linux/XFCE -- Stephen J Smoogen.
Re: Update killed samba -
On Thu, Jan 1, 2015 at 4:01 PM, Bob Goodwin bobgood...@wildblue.net wrote: I ran a system update on an SL-7 server and that seems to have messed up samba. I can ssh into the system but can't mount it as a samba share. [root@box10 bobg]# mount //192.168.1.48/myshare /mnt/box48/ Password for bobg@//192.168.1.48/myshare: Unable to find suitable address. This has worked without a problem until now. Any thoughts on what might be my problem appreciated. Which packages were updated? Did you change your samba config? Did you change your iptables config? Did you reboot after the updates? Do the client or server logs give you more info? Does smbclient -L ... list the share(s)?
Re: Update killed samba -
On Fri, Jan 2, 2015 at 7:33 AM, Bob Goodwin bobgood...@wildblue.net wrote: Jan 02 10:18:12 box10 kernel: CIFS VFS: cifs_mount failed w/return code = -113 According to /usr/include/asm-generic/errno.h, return code -113 means /* No route to host */. Akemi
Re: Update killed samba -
On Fri, Jan 2, 2015 at 3:12 PM, Bob Goodwin bobgood...@wildblue.net wrote: Only the servers are blocked from the internet since I figured they didn't require a connection but now the question of time synchronization arises. In fact my internet connection is derived from a satellite connection which has a system delay on the order of 800 ms. I would think that would offset me from the rest of the world by nearly one second if it matters ... Nahhh. Unless 'chrony' has completely mucked up the NTP spec, it deals with symmetrical propagation delays pretty well. It effectively reocords when you sent the request, when the response arrived, and what the server thought the time was to derive the relevant skews.
Re: Update killed samba -
On 01/02/15 17:08, Nico Kadel-Garcia wrote: my internet connection is derived from a satellite connection which has a system delay on the order of 800 ms. I would think that would offset me from the rest of the world by nearly one second if it matters ... Nahhh. Unless 'chrony' has completely mucked up the NTP spec, it deals with symmetrical propagation delays pretty well. It effectively reocords when you sent the request, when the response arrived, and what the server thought the time was to derive the relevant skews. Ok, that makes sense to me. I hadn't considered that ... so much to know! Bob -- http://www.qrz.com/db/W2BOD box10 Fedora-21/64bit Linux/XFCE
