Hi Philip, I tried IPs instead of hostnames at the beginning, but it didn't help: [root at host1 ~]# gluster peer probe 192.168.1.193 Probe on localhost not needed My assumption is smth wrong smwhere else. Donno where - no ideas at all; the worst is i have a set of test machines, where everything works perfectly... :( There are two switches between these machines, but they allow Gluster traffic responsively. Hope, we can struggle this together ;-) Thanks, guys, for quick responses. Regards, vy On Mon, Jul 2, 2012 at 6:27 PM, Philip Poten <philip.poten at gmail.com> wrote: > Hi Vladimir > > To circumvent any remaining /etc/hosts abnormalities, try with IPs > only. if that doesn't work either, you have a problem outside name > resolution. If it works with IPs, the problem is the resolver. > > Note that nscd caches answers, also the "host" command uses DNS > responses, /not/ the /etc/hosts (nsswitch.conf order resp.) file - but > gluster does, as does "ping" and anything else using gethostbyname(3) > > hth, > Philip > > 2012/7/2 Vladimir Yakovlev <mrquesty at gmail.com>: > > Thanks, Brian, for the hint. > > I've changed /etc/hosts with respect to your comment, but it didn't help > > either. > > > > The problem (from my perspective) in smth else, e.g. why, when I try to > do > > the following, I see the blank response in tcpdump: > > [root at host1 ~]# ip a > > 1: lo: <LOOPBACK,UP,LOWER_UP> mtu 16436 qdisc noqueue state UNKNOWN > > link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 > > inet 127.0.0.1/8 scope host lo > > 2: eth0: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc pfifo_fast > state > > DOWN qlen 1000 > > link/ether 00:25:90:30:34:42 brd ff:ff:ff:ff:ff:ff > > inet 46.../27 brd 46.182.25.159 scope global eth0 > > 3: eth1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast > state > > UP qlen 1000 > > link/ether 00:25:90:30:34:43 brd ff:ff:ff:ff:ff:ff > > inet 192.168.1.192/24 brd 192.168.1.255 scope global eth1 > > [root at host1 ~]# tcpdump -i eth1 'port 27004' > > tcpdump: verbose output suppressed, use -v or -vv for full protocol > decode > > listening on eth1, link-type EN10MB (Ethernet), capture size 65535 bytes > > ^Z > > [2]+ Stopped tcpdump -i eth1 'port 27004' > > [root at host1 ~]# bg > > [2]+ tcpdump -i eth1 'port 27004' & > > > > [root at host1 ~]# gluster peer probe host2 > > Probe on localhost not needed > > [root at host1 ~]# fg > > tcpdump -i eth1 'port 27004' > > ^C > > 0 packets captured > > 117 packets received by filter > > 52 packets dropped by kernel > > > > So (by whatever reason) Gluster doesn't send a packet to anywhere through > > ethernet. > > > > Guys, any ideas? > > > > Thanks, > > BR, > > vy > > > > > > On Mon, Jul 2, 2012 at 6:06 PM, Brian Candler <B.Candler at pobox.com> > wrote: > >> > >> On Mon, Jul 02, 2012 at 05:54:46PM +0400, Vladimir Yakovlev wrote: > >> > I tried different configurations, the latest follows: > >> > [root at host1 ~]# cat /etc/hosts > >> > 127.0.0.1 localhost localhost.localdomain localhost4 > >> > localhost4.localdomain4 > >> > 127.0.0.1 host1 host1.fqdn > >> > 192.168.1.193 host2 host2.fqdn > >> > [root at host2 ~]# cat /etc/hosts > >> > 127.0.0.1 localhost localhost.localdomain localhost4 > >> > localhost4.localdomain4 > >> > 127.0.0.1 host2 host2.fqdn > >> > 192.168.1.192 host1 host1.fqdn > >> > >> Be consistent. Both machines should have: > >> > >> 127.0.0.1 localhost localhost.localdomain > >> 192.168.1.192 host1.fqdn host1 > >> 192.168.1.193 host2.fqdn host2 > >> > >> On host1 you should have "host1.fqdn" in /etc/hostname, and the command > >> "hostname" should show it. (Ditto for host2, but with "host2.fqdn" of > >> course) > > > > > > > > _______________________________________________ > > Gluster-users mailing list > > Gluster-users at gluster.org > > http://gluster.org/cgi-bin/mailman/listinfo/gluster-users > > > -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://gluster.org/pipermail/gluster-users/attachments/20120702/44ff4a39/attachment-0001.htm>
