在Heartbeat 2.0.7版本中，如果启用了crm，则会出现如下报错
ccm[22165]: 2011/08/30_15:18:29 ERROR: REASON: can't send message to IPC: Success
cib[22166]: 2011/08/30_15:18:29 WARN: validate_cib_digest:io.c No on-disk digest present
ccm[22165]: 2011/08/30_15:18:29 ERROR: Initialization failed. Exit
cib[22166]: 2011/08/30_15:18:29 info: readCibXmlFile: [on-disk] <cib admin_epoch="0" epoch="0" num_updates="0" have_quorum="false">
heartbeat[22155]: 2011/08/30_15:18:29 WARN: Exiting /usr/lib64/heartbeat/ccm process 22165 returned rc 1.
cib[22166]: 2011/08/30_15:18:29 info: readCibXmlFile: [on-disk] <configuration>
heartbeat[22155]: 2011/08/30_15:18:29 ERROR: Respawning client "/usr/lib64/heartbeat/ccm":
heartbeat[22155]: 2011/08/30_15:18:28 WARN: G_SIG_dispatch: Dispatch function for SIGCHLD took too long to execute: 150 ms (> 10 ms) (GSource: 0x156837f8)
heartbeat[22155]: 2011/08/30_15:18:28 WARN: duplicate client add request [ccm] [22165]
heartbeat[22155]: 2011/08/30_15:18:28 ERROR: api_process_registration_msg: cannot add client()
有英文网站解释如下，处于笔者英文水平有限，无法准确翻译，请读者自己理解
This is a bug in the heartbeat API library. I'm pretty sure it has an implicit assumption that no child will connect via the heartbeat API if its parent connected before the fork, and there was no intervening exec() call...
具体网址：