X-Git-Url: http://git.onelab.eu/?a=blobdiff_plain;f=monitor%2Fbootman.py;h=8dda34116c827ee287bdb935a8e8553a216f9d30;hb=13b93b7152cc4789c9554ad11a2e1ffdd34a1304;hp=531f883e6adad05612e458f1235a7f5395643662;hpb=5772ce036b96297a23f834ea34ce4466ef4d522c;p=monitor.git diff --git a/monitor/bootman.py b/monitor/bootman.py index 531f883..8dda341 100755 --- a/monitor/bootman.py +++ b/monitor/bootman.py @@ -68,18 +68,34 @@ class NodeConnection: return "unknown" def get_dmesg(self): + t_stamp = time.strftime("%Y-%m-%d-%H:%M") self.c.modules.os.system("dmesg > /var/log/dmesg.bm.log") - download(self.c, "/var/log/dmesg.bm.log", "log/dmesg.%s.log" % self.node) - log = open("log/dmesg.%s.log" % self.node, 'r') + download(self.c, "/var/log/dmesg.bm.log", "%s/history/%s-dmesg.%s.log" % (config.MONITOR_BOOTMANAGER_LOG, t_stamp, self.node)) + os.system("cp %s/history/%s-dmesg.%s.log %s/dmesg.%s.log" % (config.MONITOR_BOOTMANAGER_LOG, t_stamp, self.node, config.MONITOR_BOOTMANAGER_LOG, self.node)) + log = open("%s/dmesg.%s.log" % (config.MONITOR_BOOTMANAGER_LOG, self.node), 'r') return log def get_bootmanager_log(self): - download(self.c, "/tmp/bm.log", "log/bm.%s.log.gz" % self.node) - #os.system("zcat log/bm.%s.log.gz > log/bm.%s.log" % (self.node, self.node)) - os.system("cp log/bm.%s.log.gz log/bm.%s.log" % (self.node, self.node)) - log = open("log/bm.%s.log" % self.node, 'r') + t_stamp = time.strftime("%Y-%m-%d-%H:%M") + download(self.c, "/tmp/bm.log", "%s/history/%s-bm.%s.log" % (config.MONITOR_BOOTMANAGER_LOG, t_stamp, self.node)) + os.system("cp %s/history/%s-bm.%s.log %s/bm.%s.log" % (config.MONITOR_BOOTMANAGER_LOG, t_stamp, self.node, config.MONITOR_BOOTMANAGER_LOG, self.node)) + log = open("%s/bm.%s.log" % (config.MONITOR_BOOTMANAGER_LOG, self.node), 'r') return log + +# def get_dmesg(self): +# self.c.modules.os.system("dmesg > /var/log/dmesg.bm.log") +# download(self.c, "/var/log/dmesg.bm.log", "log/dmesg.%s.log" % self.node) +# log = open("log/dmesg.%s.log" % self.node, 'r') +# return log +# +# def get_bootmanager_log(self): +# download(self.c, "/tmp/bm.log", "log/bm.%s.log.gz" % self.node) +# #os.system("zcat log/bm.%s.log.gz > log/bm.%s.log" % (self.node, self.node)) +# os.system("cp log/bm.%s.log.gz log/bm.%s.log" % (self.node, self.node)) +# log = open("log/bm.%s.log" % self.node, 'r') +# return log + def dump_plconf_file(self): c = self.c self.c.modules.sys.path.append("/tmp/source/") @@ -430,11 +446,16 @@ class DebugInterface: "bminit-cfg-auth-getplc-update-installinit-validate-rebuildinitrd-netcfg-disk-update4-update3-update3-implementerror-update-debug-done", "bminit-cfg-auth-getplc-installinit-validate-exception-bmexceptmount-exception-noinstall-update-debug-done", "bminit-cfg-auth-getplc-update-installinit-validate-exception-bmexceptmount-exception-noinstall-update-debug-done", + "bminit-cfg-auth-getplc-update-installinit-validate-bmexceptvgscan-exception-noinstall-update-debug-validate-bmexceptvgscan-done", + "bminit-cfg-auth-getplc-update-installinit-validate-exception-noinstall-update-debug-validate-done", ]: sequences.update({n : "restart_bootmanager_rins"}) # repair_node_keys - sequences.update({"bminit-cfg-auth-bootcheckfail-authfail-exception-update-bootupdatefail-authfail-debug-done": "repair_node_keys"}) + for n in ["bminit-cfg-auth-bootcheckfail-authfail-exception-update-bootupdatefail-authfail-debug-validate-exception-done", + "bminit-cfg-auth-bootcheckfail-authfail-exception-update-bootupdatefail-authfail-debug-done", + ]: + sequences.update({n: "repair_node_keys"}) # conn.restart_node('reinstall') for n in ["bminit-cfg-auth-getplc-update-installinit-validate-rebuildinitrd-exception-chrootfail-update-debug-done", @@ -459,12 +480,14 @@ class DebugInterface: "bminit-cfg-auth-getplc-update-installinit-validate-rebuildinitrd-netcfg-update3-implementerror-nospace-update-debug-done", "bminit-cfg-auth-getplc-hardware-installinit-installdisk-installbootfs-exception-downloadfail-update-debug-done", "bminit-cfg-auth-getplc-update-installinit-validate-implementerror-update-debug-done", + "bminit-cfg-auth-getplc-exception-update-bootupdatefail-debug-done", ]: sequences.update({n: "restart_node_boot"}) # update_node_config_email for n in ["bminit-cfg-exception-nocfg-update-bootupdatefail-nonode-debug-done", "bminit-cfg-exception-update-bootupdatefail-nonode-debug-done", + "bminit-cfg-exception-update-bootupdatefail-nonode-debug-validate-exception-done", "bminit-cfg-auth-bootcheckfail-nonode-exception-update-bootupdatefail-nonode-debug-done", ]: sequences.update({n : "update_node_config_email"}) @@ -620,8 +643,12 @@ class DebugInterface: return sequence - def restore(sitehist, hostname, config=None, forced_action=None): + ret = restore_basic(sitehist, hostname, config, forced_action) + session.flush() + return ret + +def restore_basic(sitehist, hostname, config=None, forced_action=None): # NOTE: Nothing works if the bootcd is REALLY old. # So, this is the first step. @@ -632,27 +659,19 @@ def restore(sitehist, hostname, config=None, forced_action=None): if fbnode['observed_category'] == "OLDBOOTCD": print "\t...Notify owner to update BootImage!!!" - if not found_within(recent_actions, 'newbootcd_notice', 3): + if not found_within(recent_actions, 'newbootcd_notice', 3.5): sitehist.sendMessage('newbootcd_notice', hostname=hostname) print "\tDisabling %s due to out-of-date BootImage" % hostname - api.UpdateNode(hostname, {'boot_state' : 'disable'}) + api.UpdateNode(hostname, {'boot_state' : 'disabled'}) # NOTE: nothing else is possible. return True debugnode = DebugInterface(hostname) conn = debugnode.getConnection() - #print "conn: %s" % conn - #print "trying to use conn after returning it." - #print conn.c.modules.sys.path - #print conn.c.modules.os.path.exists('/tmp/source') if type(conn) == type(False): return False - #if forced_action == "reboot": - # conn.restart_node('reinstall') - # return True - boot_state = conn.get_boot_state() if boot_state != "debug": print "... %s in %s state: skipping..." % (hostname , boot_state) @@ -682,11 +701,12 @@ def restore(sitehist, hostname, config=None, forced_action=None): print "...Should investigate. Skipping node." # TODO: send message related to these errors. - if not found_within(recent_actions, 'newbootcd_notice', 3): + if not found_within(recent_actions, 'baddisk_notice', 7): + print "baddisk_notice not found recently" log=conn.get_dmesg().read() sitehist.sendMessage('baddisk_notice', hostname=hostname, log=log) - conn.set_nodestate('disable') + conn.set_nodestate('disabled') return False @@ -723,6 +743,8 @@ def restore(sitehist, hostname, config=None, forced_action=None): args['sequence'] = s args['bmlog'] = conn.get_bootmanager_log().read() args['viart'] = False + args['saveact'] = True + args['ccemail'] = True sitehist.sendMessage('unknownsequence_notice', **args) @@ -748,7 +770,8 @@ def restore(sitehist, hostname, config=None, forced_action=None): if conn.compare_and_repair_nodekeys(): # the keys either are in sync or were forced in sync. # so try to reboot the node again. - conn.restart_bootmanager('reinstall') + # TODO: why was this originally 'reinstall' instead of 'boot'?? + conn.restart_bootmanager('boot') pass else: # there was some failure to synchronize the keys. @@ -760,6 +783,8 @@ def restore(sitehist, hostname, config=None, forced_action=None): args['sequence'] = s args['bmlog'] = conn.get_bootmanager_log().read() args['viart'] = False + args['saveact'] = True + args['ccemail'] = True sitehist.sendMessage('unknownsequence_notice', **args) conn.restart_bootmanager('boot') @@ -767,7 +792,7 @@ def restore(sitehist, hostname, config=None, forced_action=None): # TODO: differentiate this and the 'nodenetwork_email' actions. elif sequences[s] == "update_node_config_email": - if not found_within(recent_actions, 'nodeconfig_notice', 3): + if not found_within(recent_actions, 'nodeconfig_notice', 3.5): args = {} args['hostname'] = hostname sitehist.sendMessage('nodeconfig_notice', **args) @@ -775,7 +800,7 @@ def restore(sitehist, hostname, config=None, forced_action=None): elif sequences[s] == "nodenetwork_email": - if not found_within(recent_actions, 'nodeconfig_notice', 3): + if not found_within(recent_actions, 'nodeconfig_notice', 3.5): args = {} args['hostname'] = hostname args['bmlog'] = conn.get_bootmanager_log().read() @@ -784,7 +809,7 @@ def restore(sitehist, hostname, config=None, forced_action=None): elif sequences[s] == "update_bootcd_email": - if not found_within(recent_actions, 'newalphacd_notice', 3): + if not found_within(recent_actions, 'newalphacd_notice', 3.5): args = {} args.update(getconf.getconf(hostname)) # NOTE: Generates boot images for the user: args['hostname'] = hostname @@ -798,17 +823,17 @@ def restore(sitehist, hostname, config=None, forced_action=None): # require either an exception "/minhw" or other manual intervention. # Definitely need to send out some more EMAIL. # TODO: email notice of broken hardware - if not found_within(recent_actions, 'baddisk_notice', 1): + if not found_within(recent_actions, 'baddisk_notice', 7): print "...NOTIFYING OWNERS OF BROKEN HARDWARE on %s!!!" % hostname args = {} args['hostname'] = hostname args['log'] = conn.get_dmesg().read() sitehist.sendMessage('baddisk_notice', **args) - conn.set_nodestate('disable') + conn.set_nodestate('disabled') elif sequences[s] == "update_hardware_email": - if not found_within(recent_actions, 'minimalhardware_notice', 1): + if not found_within(recent_actions, 'minimalhardware_notice', 7): print "...NOTIFYING OWNERS OF MINIMAL HARDWARE FAILURE on %s!!!" % hostname args = {} args['hostname'] = hostname