troubleshooting early ssh access

[bootcd.git] / initscripts / pl_netinit

diff --git a/initscripts/pl_netinit b/initscripts/pl_netinit
index d795033..b652674 100755 (executable)
--- a/initscripts/pl_netinit
+++ b/initscripts/pl_netinit
@@ -60,7 +60,7 @@ DEFAULT_NET_CONF=0
 
 function net_init_failed() {
     echo
-    echo $(date "+%H:%M:%S") " pl_netinit: network initialization failed,"
+    echo $(date "+%H:%M:%S") " pl_netinit: network initialization failed with interface $ETH_DEVICE"
     echo
     echo For forensics
     echo
@@ -71,7 +71,11 @@ function net_init_failed() {
     echo ========== ifconfig beg
     /sbin/ifconfig
     echo ========== ifconfig end
+    echo ========== ip address show beg
+    ip address show
+    echo ========== ip address show end
     echo 
+    echo $(date "+%H:%M:%S") " pl_netinit: network initialization failed with interface $ETH_DEVICE"
     echo $(date "+%H:%M:%S") " pl_netinit: shutting down machine in two hours"
     /bin/sleep 2h
     /sbin/shutdown -h now
@@ -263,7 +267,7 @@ ETH_DEVICE=
 if [[ -n "$NET_DEVICE" ]]; then
     # the user specified a mac address we should use. find the network
     # device for it.
-    NET_DEVICE=$(tr A-Z a-z <<<$NET_DEVICE)
+    NET_DEVICE=$(tr A-Z a-z <<< $NET_DEVICE)
 
     pushd /sys/class/net
     for device in *; do
@@ -305,16 +309,25 @@ fi
 # in any case, let us try to work around that by allowing some delay
 # here
 
-ALLOW=5
+# tmp: Thierry June 2015
+# on fedora 21 nodes we see this running in a context where eth0 is not known to the system
+# could be related to a dependency that we poorly describe to systemd
+# I am increasing this timeout to 2 minutes in order to check that conjecture
+ALLOW=60
 COUNTER=0
 while true; do
     if /sbin/ifconfig $ETH_DEVICE >& /dev/null; then
-       echo "pl_netinit: device present $ETH_DEVICE, proceeding (${COUNTER}s/${ALLOW}s)"
+       echo $(date "+%H:%M:%S") " pl_netinit: device present $ETH_DEVICE, proceeding (${COUNTER}s/${ALLOW}s)"
        break
     fi
-    echo $(date "+%H:%M:%S") " pl_netinit: waiting for device $ETH_DEVICE,${COUNTER}s/${ALLOW}s"
+    echo $(date "+%H:%M:%S") " pl_netinit: waiting for device $ETH_DEVICE - ${COUNTER}s/${ALLOW}s"
+    set -x
+    /sbin/ifconfig
+    journalctl -b | egrep 'eth|bnx|udev'
+    systemctl list-unit-files | grep -i network
+    set +x 
     COUNTER=$(($COUNTER+1))
-    [ $COUNTER -ge 0 ] && net_init_failed
+    [ $COUNTER -ge $ALLOW ] && net_init_failed
     sleep 1
 done
 
@@ -336,3 +349,24 @@ echo $(date "+%H:%M:%S") " pl_netinit: attempting to start networking"
 /sbin/ifconfig $ETH_DEVICE > $IFCONFIG_OUTPUT
 
 echo $(date "+%H:%M:%S") " pl_netinit: network online"
+
+echo "BEG post pl_netinit forensics"
+
+file=/root/.ssh/authorized_keys
+echo "$file"
+if ! [ -f "$file" ]; then
+    echo "!!!!!!!!!! missing $file"
+else
+    echo "---------- $file"
+    cat $file
+fi
+
+file="/etc/ssh/sshd_config"    
+if ! [ -f "$file" ]; then
+    echo "!!!!!!!!!! missing $file"
+else
+    echo '---------- egrep Pass|Auth|PAM' $file
+    egrep 'Pass|Auth|PAM' $file
+fi
+
+echo "END post pl_netinit forensics"