On Thu, Sep 15, 2005 at 02:28:53PM +0200, Norbert Langermann wrote: > Package: heartbeat > Version: 1.2.3-9sarge3 > Severity: normal > > Hi, > > after upgrading to 1.2.3-9sarge3, stonith as called from heartbeat does not > work anymore. Calling stonith directly is no problem. > I upgraded from 1.2.3-7 or earlier, I did not look it up before I upgraded > both systems :-(
That is quite curious. I wonder if this is a side effect of the security fix that was added. Would it be possible to test 1.2.3-9 from sarge, or 1.2.3-13 from unstable? I've CCed the linux-ha list, someone there might have some more ideas. > The logfiles show normal behaviour, i.e. stonith should be executed: > -------------- > Sep 14 23:38:09 tux01-1 heartbeat[16342]: info: ************************** > Sep 14 23:38:09 tux01-1 heartbeat[16342]: info: Configuration validated. > Starting heartbeat 1.2.3 > Sep 14 23:38:09 tux01-1 heartbeat[16343]: info: heartbeat: version 1.2.3 > Sep 14 23:38:09 tux01-1 heartbeat[16343]: info: Heartbeat generation: 144 > Sep 14 23:38:09 tux01-1 heartbeat[16343]: info: ucast: write socket > priority set to IPTOS_LOWDELAY on heart0 > Sep 14 23:38:09 tux01-1 heartbeat[16343]: info: ucast: bound send socket to > device: heart0 > Sep 14 23:38:09 tux01-1 heartbeat[16343]: info: ucast: bound receive socket > to device: heart0 > Sep 14 23:38:09 tux01-1 heartbeat[16343]: info: ucast: started on port 694 > interface heart0 to 172.16.1.2 > Sep 14 23:38:09 tux01-1 heartbeat[16343]: info: ucast: write socket > priority set to IPTOS_LOWDELAY on heart0 > Sep 14 23:38:09 tux01-1 heartbeat[16343]: info: ucast: bound send socket to > device: heart0 > Sep 14 23:38:09 tux01-1 heartbeat[16343]: info: ucast: bound receive socket > to device: heart0 > Sep 14 23:38:09 tux01-1 heartbeat[16343]: info: ucast: started on port 694 > interface heart0 to 172.16.1.1 > Sep 14 23:38:09 tux01-1 heartbeat[16346]: info: pid 16346 locked in memory. > Sep 14 23:38:09 tux01-1 heartbeat[16347]: info: pid 16347 locked in memory. > Sep 14 23:38:09 tux01-1 heartbeat[16348]: info: pid 16348 locked in memory. > Sep 14 23:38:09 tux01-1 heartbeat[16343]: info: pid 16343 locked in memory. > Sep 14 23:38:09 tux01-1 heartbeat[16343]: info: Local status now set to: > 'up' > Sep 14 23:38:09 tux01-1 heartbeat[16350]: info: pid 16350 locked in memory. > Sep 14 23:38:10 tux01-1 heartbeat[16349]: info: pid 16349 locked in memory. > Sep 14 23:38:29 tux01-1 heartbeat[16343]: WARN: node tux01-2: is dead > Sep 14 23:38:29 tux01-1 heartbeat[16343]: info: Local status now set to: > 'active' > Sep 14 23:38:29 tux01-1 heartbeat[16351]: info: Resetting node tux01-2 with > [RCD_SERIAL STONITH device] > Sep 14 23:38:29 tux01-1 heartbeat[16352]: info: Checking status of STONITH > device [RCD_SERIAL STONITH device] > Sep 14 23:38:29 tux01-1 heartbeat[16343]: info: Exiting STONITH-stat > process 16352 returned rc 0. > Sep 14 23:38:39 tux01-1 heartbeat[16351]: Host tux01-2 rcd_serial-reset. > Sep 14 23:38:39 tux01-1 heartbeat[16351]: info: node tux01-2 now reset. > Sep 14 23:38:39 tux01-1 heartbeat[16343]: info: Exiting STONITH tux01-2 > process 16351 returned rc 0. > Sep 14 23:38:39 tux01-1 heartbeat[16343]: info: Resources being acquired > from tux01-2. > ------------ > > ha.cf is > ------------ > # /etc/ha.d/ha.cf > > # Log facility used in syslogd > logfacility local0 > > # Log file for all non-debug messages > # logfile /var/log/ha-log > > apiauth cl_status gid=0 uid=0 > > # nodes present in cluster > node tux01-1 tux01-2 > > stonith_host tux01-1 rcd_serial tux01-2 /dev/ttyS0 rts 10000 > stonith_host tux01-2 rcd_serial tux01-1 /dev/ttyS0 rts 10000 > > # Seconds per heartbeat > keepalive 1 > > # Other side dead after X seconds without heartbeat > deadtime 10 > > ucast heart0 172.16.1.2 > ucast heart0 172.16.1.1 > > # Get resources back to their preferred host if both are available > auto_failback yes > ------------ > > Calling "stonith -t rcd_serial -p 'tux01-2 /dev/ttyS0 rts 10000' tux01-2" > works flawlessly, and I made no changes to the config after upgrading. > > Kernel is 2.4.27-ct-1, self-compiled using the c't magazine patches: > Linux tux01-1 2.4.27-ct-1 #1 Sat Jan 15 16:14:04 UTC 2005 i686 GNU/Linux > > > Package versions: > ii heartbeat 1.2.3-9sarge3 > ii libc6 2.3.2.ds1-22 > ii libglib1.2 1.2.10-9 > ii libnet1 1.1.2.1-2 > ii libpils0 1.2.3-9sarge3 > ii libstonith0 1.2.3-9sarge3 > ii python 2.3.5-2 > ii iproute 20041019-3 > ii adduser 3.63 > ii iputils-ping 20020927-2 > ii sysklogd 1.4.1-17 > ii logrotate 3.7-5 > > As the HA system is already in production use, quick help would be > appreciated :-) > Thanks, > Norbert > -- > Norbert Langermann > [EMAIL PROTECTED] > OpenPGP-Key 0x51431418 > > Verschränkung ist, wenn man das eine Teilchen kitzelt und das andere > lacht. > -- Horms
