Hi, On Tue, May 24, 2011 at 09:43:50AM +0200, Bart Coninckx wrote: > On 05/23/11 23:34, Lars Ellenberg wrote: > > On Mon, May 23, 2011 at 01:59:23PM -0700, Hai Tao wrote: > >> this might not be too close to HA, but I am not sure if someone has seem > >> this before: > >> > >> I use a serial cable between two nodes, and I am testing the heartbeat > >> with : > >> > >> > >> server2$ cat< /dev/ttyS0 > >> server1$ echo hello> /dev/ttyS0 > >> > >> instead of receiving "hello" on server2, I see some hashed code there. > >> > >> Does someone have an idea why I do not receive the "hello" in clear text? > > Mismatch of settings, especially baud, flow control, or similar. > > If you really want to do this manually, learn about stty. > > > > for starters, try, on both nodes, > > stty -F /dev/ttyS0 > > and compare the output. > > > > Then try stty 115200 cs8 -F /dev/ttyS0 > > and add whatever else you need to get useful settings. > > > > You also need a "null modem cable", usually, > > not just any serial cable. > > > > Note that, in contrast to the "haresources" mode, with Pacemaker, > > not only small heartbeats are exchanged, but larger stringified > > XML, occasionally even the whole CIB, inclusive configuration and > > status sections. > > > > Which, with a few resources, even when compressed, can reach an > > "unexpected" volume. > > > > Consider the transfer time of even only 10kByte on a serial port > > connection. (Yes, that's ~one second, on a fast port!). > > > > You want the highest possible stable baud rate, the smallest > > possible pacemaker configuration, and timeouts that take this into > > account. > > > > In my experience, boxes that have high volume serial port activity > > can feel very sluggish in all aspects. > > > > For non-haresources clusters, we recommend against serial > > communication paths. We also recommend against haresources > > clusters, unless that really is all you want and need. > > > > So probably just forget about serial communication paths, > > but use all available physically independend network links, > > then add an other two ;-) > > > > > Just out of curiosity: why the recommendation against serial > communication (as backup medium for instance) in non-haresources clusters?
Quoting Lars again: > > Note that, in contrast to the "haresources" mode, with Pacemaker, > > not only small heartbeats are exchanged, but larger stringified > > XML, occasionally even the whole CIB, inclusive configuration and > > status sections. > > > > Which, with a few resources, even when compressed, can reach an > > "unexpected" volume. Heartbeat always uses all media to transmit messages, i.e. you cannot have serial as passive interconnect. But even if that were possible it wouldn't help much with big configurations. Thanks, Dejan > thx, > > B. > > > _______________________________________________ > Linux-HA mailing list > [email protected] > http://lists.linux-ha.org/mailman/listinfo/linux-ha > See also: http://linux-ha.org/ReportingProblems _______________________________________________ Linux-HA mailing list [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
