load_checkpoint.sh #28

#!/bin/bash
#------------------------------------------------------------------------------
### WARNING WARNING WARNING WARNING WARNING WARNING WARNING WARNING WARNING
### THIS SCRIPT SHOULD BE USED WITH GREAT CAUTION.
#------------------------------------------------------------------------------
set -u

# This script loads a specified checkpoint into /p4/N/root and /p4/N/offline_db.

#==============================================================================
# Declarations and Environment

# Allow override of P4U_HOME, which is set only when testing P4U scripts.
export P4U_HOME="${P4U_HOME:-/p4/common/bin}"
export P4U_LIB="${P4U_LIB:-/p4/common/lib}"
export P4U_ENV="$P4U_LIB/p4u_env.sh"
export P4U_LOG=Unset
export VERBOSITY=${VERBOSITY:-5}

# Environment isolation.  For stability and security reasons, prepend
# PATH to include dirs where known-good scripts exist.
# known/tested PATH and, by implication, executables on the PATH.
export PATH="$P4U_HOME:$PATH:~/bin:."
export P4CONFIG="${P4CONFIG:-.p4config}"

[[ -r "$P4U_ENV" ]] || {
   echo -e "\\nError: Cannot load environment from: $P4U_ENV\\n\\n"
   exit 1
}

declare BASH_LIBS=$P4U_ENV
BASH_LIBS+=" $P4U_LIB/libcore.sh"
BASH_LIBS+=" $P4U_LIB/libp4u.sh"

for bash_lib in $BASH_LIBS; do
   # shellcheck disable=SC1090
   source "$bash_lib" ||\
      { echo -e "\\nFATAL: Failed to load bash lib [$bash_lib]. Aborting.\\n"; exit 1; }
done

declare ThisScript=${0##*/}
declare CmdLine="$0 $*"
declare ThisUser=
declare Version=2.7.0
declare -i SilentMode=0
declare -i GenSSLCertsIfNeeded=1
declare -i LicenseFileNeeded=1
declare -i StartWithoutLicenseFile=0
declare -i DoVerify=0
declare -i LoadLiveJournal=0
declare DefaultVerifyOptions="-o MISSING"
declare VerifyOptions=
declare -i i=0
declare Checkpoint=Unset
declare SDPInstance=
declare SDPInstanceVarsFile=
declare OfflineDB=
declare OfflineDBUsableFile=
declare CaseMode=
declare CaseFlag=
declare CompressFlag=
declare P4BrokerCfg=
declare P4DPids=
declare SetServerID=
declare ReplicaTypeTag=
declare InferredReplicaTypeTag=
declare ThisHost=${HOSTNAME%%.*}
declare P4DInitScript=
declare P4DSystemdServiceFile=
declare P4BrokerInitScript=
declare P4BrokerSystemdServiceFile=
declare Datestamp=
declare INode1=
declare INode2=
declare -i JournalCount=0
declare -a Journals

#==============================================================================
# Local Functions

function user_confirmation_and_warning () {
   local interactive=${1:-1}
   local input=""
   msg "\\n
==============================================================================
WARNING WARNING WARNING WARNING WARNING WARNING WARNING WARNING WARNING\\n

This script will remove files in P4ROOT and the offline_db folder,
including db.* files, state* files, and others.

It is STRONGLY recommended that these files be preserved first unless
it is absolutely certain they will never potentially be useful in any
recovery scenario. See the '-k' option.

This is being run on host $HOSTNAME at $(date).

WARNING WARNING WARNING WARNING WARNING WARNING WARNING WARNING WARNING
==============================================================================
\\n\\n"

   if [[ $interactive -eq 0 ]]; then
      msg "Bypassing interactive confirmation due to -y on command line."
      return 0
   fi

   while [[ -z "$input" ]]; do
      echo -e -n "\\nEnter Y to proceed, N to stop [y/Y/n/N]:  "
      read -r -e input
      if [[ "${input^^}" == "Y" ]]; then
         continue
      elif [[ "${input^^}" == "N" ]]; then
         msg "Confirmation to proceed not received."
         return 1
      else
         input=
      fi
   done

   return 0
}


#------------------------------------------------------------------------------
# Function: terminate
function terminate
{
   # Disable signal trapping.
   trap - EXIT SIGINT SIGTERM

   vvmsg "$ThisScript: EXITCODE: $OverallReturnStatus"

   # Stop logging.
   stoplog

   # Don't litter.
   cleanTrash

   # With the trap removed, exit.
   exit "$OverallReturnStatus"
}

#------------------------------------------------------------------------------
# Function: usage (required function)
#
# Input:
# $1 - style, either -h (for short form) or -man (for man-page like format).
# The default is -h.
#
# $2 - error message (optional).  Specify this if usage() is called due to
# user error, in which case the given message displayed first, followed by the
# standard usage message (short or long depending on $1).  If displaying an
# error, usually $1 should be -h so that the longer usage message doesn't
# obscure the error message.
#
# Sample Usage:
# usage
# usage -h
# usage -man
# usage -h "Incorrect command line usage."
#------------------------------------------------------------------------------
function usage
{
   declare style=${1:--h}
   declare errorMessage=${2:-Unset}

   if [[ "$errorMessage" != Unset ]]; then
      echo -e "\\n\\nUsage Error:\\n\\n$errorMessage\\n\\n"
   fi

   echo "USAGE for $ThisScript v$Version:

$ThisScript <checkpoint> [<jnl.1> [<jnl.2> ...]] [-k] [-i <instance>] [-s <ServerID>] [-t <Type>] [-verify {default|\"Verify Options\"}] [-c] [-l] [-r] [-b] [-y] [-L <log>] [-si] [-v<n>] [-D]

or

$ThisScript [-h|-man|-V]
"
   if [[ $style == -man ]]; then
      echo -e "
DESCRIPTION:
	This script loads a specified checkpoint into /p4/N/root and /p4/N/offline_db,
	where 'N' is the SDP instance name.

	At the start of processing, preflight checks are done. Preflight checks
	include:
	* The specified checkpoint and corresponding *.md5 file must exist.
	* All journal files to replay (if any were specified) must exist.
	* The \$P4ROOT/server.id file must exist, unless '-s' is specified.
	* If the \$P4ROOT/server.id file exists and '-s' is specified, the values
	must match.
	* The \$P4ROOT/license file must exist, unless '-l' is specified or if
	the replica type does not require a license.
	* Basic SDP structure and key files must exist.

  	If the preflight passes, the p4d_N service is shutdown, and also the
	p4broker_N service is shutdown if configured.

	If a P4LOG file exists, it is moved aside so there is a fresh p4d server
	log corresponding to operation after the checkpoint load.

	If a P4JOURNAL file exists, it is moved aside as the old journal data is
	no longer relevant after a checkpoint replay. (Exception: If the P4JOURNAL
	is speciffed in a list of journals to reply, then it is not moved aside).

	Next, any existing state* files in P4ROOT are removed.

	Next, any existing database files in P4ROOT are removed (or moved aside
	with '-k').

	Next, the specified checkpoint is loaded. Upon completion, the Helix Core
	server process, p4d_N, is started.

	If the server to be started is a replica, the serviceUser configured for
	the replica is logged into the P4TARGET server.  Any needed 'p4 trust' and
	'p4 login' commands are done to enable replication.

	Note that this part of the processing will fail if the correct super user
	password is not stored in the standard SDP password file,

	/p4/common/config/.p4passwd.p4_N.admin

	After starting the server, a local 'p4 trust' is done if needed, and then
	a 'p4login -service -v' and 'p4login -v'.

	By default, the p4d_N service is started, but the p4broker_N service is not.
	Specify '-b' to restart both services.

ARGUMENTS AND OPTIONS:
 <checkpoint>
	Specify the path to the checkpoint file to load. Exactly one checkpoint
	must be specified.

	The file may be a compressed or uncompressed checkpoint, and it may be a case
	sensitive or case-insensitive checkpoint.  The checkpoint file must have a
	corresponding *.md5 checksum file in the same directory, with one of two name
	variations: If the checkpoint file is /somewhere/foo.gz, the checksum file may
	be named /somewhere/foo.gz.md5 or /somewhere/foo.md5.

<jnl.1> [<jnl.2> ...]
	Specify the path to the one or more journal files to replay after the
	checkpoint, in the correct sequence order.

 -k	Specify '-k' to keep db.* files in P4ROOT rather than removing them.  This may
	be desirable to preserve databases for investigation.  This option requires
	sufficient disk space to hold an extra copy of the db.* files.

	If -k specified, a folder named 'MovedDBs.<datestamp>' is created under the
	P4ROOT directory, and databases are moved there.

 -i <instance>
	Specify the SDP instance. This can be omitted if SDP_INSTANCE is already
	defined.

 -s <ServerID>
	Specify the ServerID.  This value is written into \$P4ROOT/server.id file.

	If no \$P4ROOT/server.id file exists, this flag is required.

	If the \$P4ROOT/server.id file exists, this argument is not needed. If this
	'-s <ServerID>' is given and a \$P4ROOT/server.id file exists, the value in
	the file must match the value specified with this argument.

 -t <Type>
	Specify the replica type tag if the checkpoint to be loaded is for an edge
	server or replica. The set of valid values for the replica type
	tag are defined in the documentation for mkrep.sh.  See: mkrep.sh -man

	If the type is specified, the '-s <ServerID>' is required.

	If the SDP Server Spec Naming Standard is followed, the ServerID
	specified with '-s' will start with 'p4d_'. In that case, the
	value for '-t edge' value is inferred, and '-t' is not required.

	If the type is specified or inferred, certain behaviors change based
	on the type:
	* If the type is edge, only the correct edge-specific subset of database
	tables are loaded.
	* The P4ROOT/license file check is suppressed unless the type is
	ha, ham, fs, for fsm (standby replicas usable with 'p4 failover').

	Do not use this '-t <Type>' option if the checkpoint being loaded is
	for a master server.

	For an edge server, an edge seed checkpoint created with edge_dump.sh
	must be used if the edge is filtered, e.g. if any of the *DataFilter
	fields in the server spec are used. If the edge server is not filtered
	by means other than being an edge server (for which certain tables are
	filtered by nature), a standard full checkpoint from the master can be
	used.

	For a filtered forwarding replica, a proper seed checkpoint must be
	loaded.  This can be created on the master using key options to p4d,
	including '-P <ServerID> -jd <SeedCkp' on the master (possibly using
	the 'offline_db' to avoid downtime, similar to how edge_dump.sh
	works for edge servers).

	WARNING: While this script is useful for seeding a new edge server, this
	script is NOT to be used for recovering or reseeding an existing edge server,
	because all edge-local database tables (mostly workspace data) would be lost.
	To recover an existing edge server, see the recover_edge.sh script.

	Warning: If this option is specified with the incorrect type for the
	checkpoint specified, results will be unpredictable.

 -verify default
 -verify \"Verify Options\"

	Specify '-verify' to initiate a call to 'p4verify.sh' after the server
	is online. On a replica, this can be useful to cause the server to pull
	missing archive files from its P4TARGET server. If this load_checkpoint.sh
	script is used in a recovery situation for a master server, this '-verify'
	option can be used to discover if archive files are missing after the
	metadata is recovered.

	The 'p4verify.sh' script has a rich set of options; see 'p4verify.sh -man'
	for more info.  To options to pass to p4verify.sh can be passed in a
	quoted list, or the value 'verify default' can be used to indicate these
	options:

	$DefaultVerifyOptions

	By default, a fast verify is used if the p4d version is new enough (2021.1+).
	See 'p4verify.sh -man' for more information, specifically the description
	of the '-o MISSING' option.

	In all cases, p4verify.sh is invoked as a background process; this
	load_checkpoint.sh script does not wait for it to complete. The p4verify.sh
	script will email as per normal when it completes.

 -c	Specify that SSL certificates are required, and not to be generated with
	'p4d_N -Gc'.

	By default, if '-c' is not supplied and SSL certs are not available, certs
	are generated automatically with 'p4d_N -Gc'.

 -l	Specify that the server is to start without a license file. By default, if
	there is no \$P4ROOT/license file, this script will abort.  Note that if '-l'
	is specified and a license file is actually needed, the attempt this script makes
	to start the server after loading the checkpoint will fail.

	If -'t <type>' is specified, the license check is skipped unless the type is
	'ha' or 'ham', i.e. HA replicas that need a license file to support failover.

 -r	Specify '-r' to replay only to P4ROOT.  By default, this script replays both
	to P4ROOT and the offline_db.

 -b	Specify '-b' to start the a p4broker process (if configured).  By default
	the p4d process is started after loading the checkpoint, but the p4broker
	process is not. This can be useful to ensure the human administrator has
	an opportunity to do sanity checks before enabling the broker to allow
	access by end users (if the broker is deployed for this usage).

 -y	Use the '-y' flag to bypass an interactive warning and confirmation
	prompt.

 -v<n>	Set verbosity 1-5 (-v1 = quiet, -v5 = highest).  The default is 5.

 -L <log>
	Specify the path to a log file.  By default, all output (stdout and stderr)
	goes to:
	/p4/<instance>/logs/${ThisScript%.sh}.<timestamp>.log

	NOTE: This script is self-logging.  That is, output displayed on the screen
	is simultaneously captured in the log file.  Do not run this script with
	redirection operators like '> log' or '2>&1', and do not use 'tee.'

 -si	Operate silently.  All output (stdout and stderr) is redirected to the log
	only; no output appears on the terminal.
      
 -D     Set extreme debugging verbosity.

HELP OPTIONS:
 -h	Display short help message
 -man	Display man-style help message
 -V	Display version info for this script and its libraries.

EXAMPLES:
	EXAMPLE 1: Non-interactive Usage

	Non-interactive usage (bash syntax) to load a checkpoint:

	nohup $P4CBIN/load_checkpoint.sh /p4/1/checkpoints/p4_1.ckp.4025.gz -i 1 -y < /dev/null > /dev/null 2>&1 &

	Then, monitor with:
	tail -f \$(ls -t \$LOGS/load_checkpoint.*.log|head -1)\\n

	EXAMPLE 2: Checkpoint Load then Verify

	Non-interactive usage (bash syntax) to load a checkpoint followed by a full
	verify of recent archives files only with other options passed to verify.sh:

	nohup $P4CBIN/load_checkpoint.sh /p4/1/checkpoints/p4_1.ckp.4025.gz -i 1 -verify "-recent -nu -ns" -y < /dev/null > /dev/null 2>&1 &

	EXAMPLE 3: Load Checkpoint and Journals

	Non-interactive usage (bash syntax) to loading a checkpoint and subsequent journals:

	nohup $P4CBIN/load_checkpoint.sh /p4/1/checkpoints/p4_1.ckp.4025.gz /p4/1/checkpoints/p4_1.jnl.4025 /p4/1/checkpoints/p4_1.jnl.4026 -i 1 -y < /dev/null > /dev/null 2>&1 &

	Then, monitor with:
	tail -f \$(ls -t \$LOGS/load_checkpoint.*.log|head -1)\\n

	EXAMPLE 4: Interactive usage.

	Interactive usage to load a checkpoint with no license file.

	$P4CBIN/load_checkpoint.sh /p4/1/checkpoints/p4_1.ckp.4025.gz -i 1 -l

	With interactive usage, logging still occurs; all output to the screen is
	captured.

	Note that non-interactive usage with nohup is recommended for checkpoints
	with a long replay duration, to make operation more reliable in event of a
	shell session disconnect.  Alternately, running interactively in a 'screen'
	session (if 'screen' is available) provides similar protection against
	shell session disconnects.

	EXAMPLE 5: Seed New Edge

	Seeding a new edge server.

	nohup $P4CBIN/load_checkpoint.sh /p4/1/checkpoints/p4_1.ckp.4025.gz -i 1 -s p4d_edge_syd < /dev/null > /tmp/null 2>&1 &

	WARNING: While this script is useful for seeding a new edge server, this
	script is NOT to be used for recovering or reseeding an existing edge server,
	because all edge-local database tables (mostly workspace data) would be lost.
	To recover an existing edge server, see the recover_edge.sh script.

	EXAMPLE 6: Seed New Edge and Verify

	Seeding a new edge server and then do a verify with default options.

	nohup $P4CBIN/load_checkpoint.sh /p4/1/checkpoints/p4_1.ckp.4025.gz -i 1 -s p4d_edge_syd -verify default < /dev/null > /tmp/null 2>&1 &
"
   fi

   exit 1
}

#==============================================================================
# Command Line Processing

declare -i LoadOfflineDB=1
declare -i StartBrokerWhenDone=0
declare -i Interactive=1
declare -i KeepRootDBs=0
declare -i shiftArgs=0

set +u
while [[ $# -gt 0 ]]; do
   case $1 in
      (-i) SDPInstance="$2"; shiftArgs=1;;
      (-s) SetServerID="$2"; shiftArgs=1;;
      (-t) ReplicaTypeTag="$2"; shiftArgs=1;;
      (-verify) DoVerify=1; VerifyOptions="$2" shiftArgs=1;;
      (-c) GenSSLCertsIfNeeded=0;;
      (-k) KeepRootDBs=1;;
      (-l) StartWithoutLicenseFile=1;;
      (-r) LoadOfflineDB=0;;
      (-b) StartBrokerWhenDone=1;;
      (-y) Interactive=0;;
      (-h) usage -h;;
      (-man) usage -man;;
      (-V) show_versions; exit 1;;
      (-v1) export VERBOSITY=1;;
      (-v2) export VERBOSITY=2;;
      (-v3) export VERBOSITY=3;;
      (-v4) export VERBOSITY=4;;
      (-v5) export VERBOSITY=5;;
      (-L) export P4U_LOG="$2"; shiftArgs=1;;
      (-si) SilentMode=1;;
      (-D) set -x;; # Debug; use 'set -x' mode.
      (-*) usage -h "Unknown flag ($1).";;
      (*)
         if [[ "$Checkpoint" == Unset ]]; then
            Checkpoint="$1"
         else
            Journals[$JournalCount]="$1"
            JournalCount+=1
         fi
      ;;
   esac

   # Shift (modify $#) the appropriate number of times.
   shift; while [[ $shiftArgs -gt 0 ]]; do
      [[ $# -eq 0 ]] && usage -h "Incorrect number of arguments."
      shiftArgs=$shiftArgs-1
      shift
   done
done
set -u

#==============================================================================
# Command Line Verification

[[ "${P4U_LOG:-Unset}" == Unset ]] && \
   P4U_LOG="${LOGS:-/tmp}/${ThisScript%.sh}.$(date +'%Y%m%d-%H%M%S').log"

if [[ -z "$SDPInstance" ]]; then
   if [[ "${SDP_INSTANCE:-Unset}" != Unset ]]; then
      SDPInstance="$SDP_INSTANCE"
   else
      usage -h "SDP Instance parameter is missing and SDP_INSTANCE is undefined." 
   fi
fi

if [[ -n "$ReplicaTypeTag" && -z "$SetServerID" ]]; then
   usage -h "The '-t' was used without '-s'.  If '-t <Type>' is specified, the '-s <ServerID>' argument is required."
fi

if [[ -n "$VerifyOptions" && "$VerifyOptions" == default ]]; then
   VerifyOptions="$DefaultVerifyOptions"
fi

#==============================================================================
# Main Program

trap terminate EXIT SIGINT SIGTERM

declare -i OverallReturnStatus=0

# Initialize logging.
touch "${P4U_LOG}" || bail "Couldn't touch log file [${P4U_LOG}]."

# Redirect stdout and stderr to a log file.
if [[ "$SilentMode" -eq 0 ]]; then
   exec > >(tee "${P4U_LOG}")
   exec 2>&1
else
   exec >"${P4U_LOG}"
   exec 2>&1
fi

initlog

ThisUser=$(id -n -u)
msg "Started $ThisScript v$Version as $ThisUser@$ThisHost on $(date) as:\\n$CmdLine\\n"

# shellcheck disable=SC2153
SDPInstanceVarsFile="/p4/common/config/p4_${SDPInstance}.vars"

[[ "$Checkpoint" == Unset ]] && usageError "Missing checkpoint parameter."
[[ -r "$SDPInstanceVarsFile" ]] ||\
   bail "Missing SDP instance vars file. Typo in instance name? Expected file: $SDPInstanceVarsFile"
[[ -r "$Checkpoint" ]] ||\
   bail "Specified checkpoint file does not exist: $Checkpoint"

# Preflight - check that specified journals exist. Also, determine if the live
# P4JOURNAL file is on the list of journal to be replied.
if [[ "$JournalCount" -gt 0 ]]; then
   if [[ -r "$P4JOURNAL" ]]; then
      # shellcheck disable=SC2012
      INode1="$(ls -i "$P4JOURNAL" 2>/dev/null|awk '{print $1}')"
   fi

   for jnl in "${Journals[@]}"; do
      if [[ -r  "$jnl" ]]; then
         if [[ -n "$INode1" ]]; then
            # shellcheck disable=SC2012
            INode2="$(ls -i "$jnl" 2>/dev/null|awk '{print $1}')"
            [[ "$INode1" == "$INode2" ]] && LoadLiveJournal=1
         fi
      else
         bail "Specified journal file does not exist: $jnl"
      fi
   done
fi

# shellcheck disable=SC1091
source /p4/common/bin/p4_vars "$SDPInstance"
# shellcheck disable=SC1091
source /p4/common/bin/ps_functions.sh ||\
   bail "Failed to load ps_functions.sh."
#shellcheck disable=SC1091
source /p4/common/bin/edge_vars ||\
   bail "Failed to load edge_vars."
declare CheckpointMD5=

# Set or check ServerID value.
if [[ -n "$SetServerID" ]]; then
   if [[ -f "$P4ROOT/server.id" ]]; then
      if grep -q "$SetServerID" "$P4ROOT/server.id"; then
         msg "Verified: $P4ROOT/server.id value matches set value: $SetServerID"
      else
         bail "Existing $P4ROOT/server.id file is not set to $SetServerID, but to $(cat "$P4ROOT"/server.id). If you are certain you want to use the ServerID of $SetServerID specified with '-s', remove the existing $P4ROOT/server.id file, e.g. with:\\n\\trm -f $P4ROOT/server.id\\n\\nAnd then try again. Aborting for now."
      fi
   else
      msg "Setting ServerID $SetServerID in $P4ROOT/server.id"
      echo "$SetServerID" > "$P4ROOT/server.id" ||\
         bail "Failed to write $P4ROOT/server.id. Aborting"
   fi
fi

# Infer the replica type from the ServerID if possible (if the SDP Server Spec
# Naming Standard was followed).
if [[ -n "$SetServerID" && "$SetServerID" == "p4d_"* ]]; then
   # To get the type tag, trim the 'p4d_' prefix ...
   InferredReplicaTypeTag=${SetServerID#p4d_}
   # then trim the "_<SiteTag>" suffix ...
   InferredReplicaTypeTag=${InferredReplicaTypeTag%%_*}
   # and finally remove numbers, as the naming standard allows
   # for things like p4d_ha2_nyc or p4d_edge2_syd, and we don't
   # want the numbers in the type tag.
   InferredReplicaTypeTag=${InferredReplicaTypeTag/[0-9]*/}
   msg "Inferred '-t $InferredReplicaTypeTag' from '-s $SetServerID'."

   if [[ -z "$ReplicaTypeTag" ]]; then
      ReplicaTypeTag="$InferredReplicaTypeTag"
   elif [[ "$ReplicaTypeTag" == "$InferredReplicaTypeTag" ]]; then
      msg "Verified: Value specified with '-t $InferredReplicaTypeTag' matches value inferred from '-s $SetServerID'."
   else
      usage -h "The replica type '-t $ReplicaTypeTag' was specified but does not match the value '$InferredReplicaTypeTag' inferred from '-s $SetServerID'."
   fi
fi

# The replica type may have been set with '-t <Type>' or inferred from
# '-s <ServerID>'. In either case, disable the license check if it's
# not needed.
if [[ -n "$ReplicaTypeTag" ]]; then
   # If a replica, the license file is not required by default unless
   # the type tag indicates it is a type used for High Availability,
   # which should have a license file to support 'p4 failover'.
   case "$ReplicaTypeTag" in
      (ha|ham) true;;
      (fs|fsm) true;;
      (*) LicenseFileNeeded=0;;
   esac
fi

# shellcheck disable=SC2153
P4DInitScript="$P4HOME/bin/p4d_${SDP_INSTANCE}_init"
P4DSystemdServiceFile="/etc/systemd/system/p4d_${SDP_INSTANCE}.service"
P4BrokerInitScript="$P4HOME/bin/p4broker_${SDP_INSTANCE}_init"
P4BrokerSystemdServiceFile="/etc/systemd/system/p4broker_${SDP_INSTANCE}.service"

user_confirmation_and_warning "$Interactive" ||\
   bail "User confirmation not verified. Aborting."

cd "$P4ROOT" || bail "Could not cd to P4ROOT [$P4ROOT]."

[[ "$PWD" == "/p4/${SDPInstance}/root" ]] || bail "Unexpected P4ROOT value of: $P4ROOT"

CheckpointMD5="${Checkpoint}.md5"
if [[ -r "$CheckpointMD5" ]]; then
   msg "Verified: MD5 file for checkpoint exists: $CheckpointMD5"
else
   CheckpointMD5="${Checkpoint/.gz/.md5}"
   if [[ -r "$CheckpointMD5" ]]; then
      msg "Verified: MD5 file for checkpoint exists: $CheckpointMD5"
   else
      bail "Could not find MD5 file for checkpoint. Tried ${Checkpoint}.md5 and ${Checkpoint/.gz/.md5}"
   fi
fi

if [[ "$LicenseFileNeeded" -eq 1 ]]; then
   msg "Checking for license file."
   if [[ ! -r "$P4ROOT/license" ]]; then
      if [[ "$StartWithoutLicenseFile" -eq 1 ]]; then
         warnmsg "No $P4ROOT/license file found. Continuing due to '-l'."
      else
         bail "No $P4ROOT/license file found. If none needed, specify '-l' and try again."
      fi
   fi
fi

if [[ "$P4PORT" =~ ^ssl[46]*: && ! -r "$P4SSLDIR/certificate.txt" ]]; then
   if [[ "$GenSSLCertsIfNeeded" -eq 1 ]]; then
      if [[ ! -d "$P4SSLDIR" ]]; then
         mkdir -p "$P4SSLDIR" ||\
            bail "Failed to create P4SSLDIR [$P4SSLDIR]."
      fi

      chmod 700 "$P4SSLDIR" ||\
         bail "Failed to do 'chmod 700 $P4SSLDIR'."

      "$P4DBIN" -Gc ||\
         bail "There was a problem generating SSL certificates with '$P4DBIN -Gc'."
   else
      bail "The P4PORT has SSL is enabled, but there is no certificate file\\n[$P4SSLDIR/certificate.txt], and the '-c' flag as specified requiring SSL certificates to be in place.\\n"
   fi
fi

[[ -x "$P4DInitScript" ]] || bail "Aborting: P4D Init script is missing or not executable: $P4DInitScript"

msg "Checking to see if broker is configured."
if [[ -r "/p4/common/config/p4_${SDPInstance}.broker.${ThisHost}.cfg" ]]; then
   P4BrokerCfg="/p4/common/config/p4_${SDPInstance}.broker.${ThisHost}.cfg"
   msg "Broker configured with: $P4BrokerCfg"
elif [[ -r "/p4/common/config/p4_${SDPInstance}.broker.cfg" ]]; then
   P4BrokerCfg="/p4/common/config/p4_${SDPInstance}.broker.cfg"
   msg "Broker configured with: $P4BrokerCfg"
else
   msg "No broker config detected."
fi

if [[ -x "$P4BrokerInitScript" && -r "$P4BrokerCfg" ]]; then
   msg "Checking p4broker status."
   if "$P4BrokerInitScript" status; then
      msg "Shutting down p4broker."
      if [[ -r "$P4BrokerSystemdServiceFile" ]]; then
         sudo systemctl stop "${P4BROKERBIN##*/}" ||\
            bail "Failed to execute: sudo systemctl stop ${P4BROKERBIN##*/}"
      else
         "$P4BrokerInitScript" stop
      fi
   else
      msg "Verified: p4broker is down."
   fi
fi

msg "Checking p4d status."
P4DPids=$(get_pids "$P4DBIN")

if [[ -n "$P4DPids" ]]; then
   msg "Shutting down p4d."
   if [[ -r "$P4DSystemdServiceFile" ]]; then
      sudo systemctl stop "${P4DBIN##*/}" ||\
         bail "Failed to execute: sudo systemctl stop ${P4DBIN##*/}"
   else
      "$P4DInitScript" stop
   fi
else
   msg "Verified: No processes are running for: $P4DBIN"
fi

echo "P4ROOT is not available during load_checkpoint.sh processing." > "$P4ROOT/P4ROOT_not_usable.txt"

# Remove state* files.
Cmd="/bin/rm -f state*"
msg "Cleaning up state* files with this command in $PWD:\\n$Cmd"
$Cmd || bail "Failed to clean up old state files in $P4ROOT"

# Move aside or Remove db.* files, but not server.id or license files.
Datestamp=$(date +'%Y%m%d-%H%M')
if [[ "$KeepRootDBs" -eq 1 ]]; then
   MovedDBDir="MovedDBs.$Datestamp"
   mkdir "$MovedDBDir" || bail "In $PWD could not do: mkdir \"$MovedDBDir\""
   Cmd="/bin/mv -f db.* $MovedDBDir/."
   msg "Moving aside databases with this command in $PWD:\\n$Cmd"
   $Cmd || bail "Failed to move aside up old databases in $P4ROOT"
else
   Cmd="/bin/rm -f db.*"
   msg "Cleaning up databases with this command in $PWD:\\n$Cmd"
   $Cmd || bail "Failed to clean up old databases in $P4ROOT"
fi

# Move P4LOG aside if it exists.
if [[ -r "$P4LOG" ]]; then
   Cmd="mv $P4LOG $P4LOG.moved.$(date +'%Y%m%d-%H%M')"
   msg "Moving P4LOG aside with this command:\\n$Cmd"
   $Cmd || bail "Failed to move P4LOG aside."
fi

# Move P4JOURNAL aside if it exists, unless the live P4JOURNAL was on
# the list of journal to replay.
if [[ "$LoadLiveJournal" -eq 0 && -r "$P4JOURNAL" ]]; then
   Cmd="mv $P4JOURNAL $P4JOURNAL.moved.$(date +'%Y%m%d-%H%M')"
   msg "Moving P4JOURNAL aside with this command:\\n$Cmd"
   $Cmd || bail "Failed to move P4JOURNAL aside."
fi

# Move journal.NNN files (for standby replicas) aside if they exist.
cd "$LOGS" || bail "Could not cd to: $LOGS"

# shellcheck disable=SC2045
for f in $(ls journal.* 2>/dev/null); do
   [[ "$f" =~ ^journal.[0-9]+$ ]] || continue
   Cmd="mv $f $f.moved.$Datestamp"
   msg "Moving file $f aside with this command:\\n$Cmd"
   $Cmd || bail "Failed to move file $f aside."
done
cd - > /dev/null || bail "Could not cd to: $OLDPWD"

# For checkpoint replay operations, determine whether case sensitivity
# flag is needed, and whether '-z' is needed.
if [[ "$Checkpoint" == *".gz" ]]; then
   if [[ "$(uname -s)" == "Darwin" ]]; then
      CaseMode=$(zcat < "$Checkpoint" | head -1 | grep -E '^@nx@ (0|2) ' | cut -d ' ' -f5)
   else
      CaseMode=$(zcat "$Checkpoint" | head -1 | grep -E '^@nx@ (0|2) ' | cut -d ' ' -f5)
   fi

   if [[ "$CaseMode" == "2" ]]; then
      CaseFlag='-C1'
      msg "Case-insensitive checkpoint detected."
   else
      msg "Case-sensitive checkpoint detected."
   fi

   # In P4D 2018.1, the '-z' flag should be dropped, as the need for
   # compression/decompression is determined by P4D when replying,
   # and thus '-z' should not be used (though it is supported by P4D
   # for backward compatibility, and we use it here if using an older
   # server.
   # shellcheck disable=SC2072
   [[ "$P4D_VERSION" < "2018.1" ]] && CompressFlag='-z'
else
   CaseMode=$(head -1 "$Checkpoint"| grep -E '^@nx@ (0|2) ' | cut -d ' ' -f5)
   if [[ "$CaseMode" == "2" ]]; then
      CaseFlag='-C1'
      msg "Case-insensitive checkpoint detected."
   else
      msg "Case-sensitive checkpoint detected."
   fi
fi

Cmd="$P4DBIN -r $P4ROOT $CompressFlag $CaseFlag"

# For edge servers, ignore tables specified by $ExcludedTables, defined in
# /p4/common/bin/edge_vars.
if [[ "$ReplicaTypeTag" == "edge" ]]; then
   # shellcheck disable=SC2154
   Cmd+=" -K $ExcludedTables"
fi

Cmd+=" -jr $Checkpoint"

msg "Replaying checkpoint to P4ROOT with this command:\\n$Cmd"
$Cmd || bail "Checkpoint replay to P4ROOT failed."

if [[ "$JournalCount" -gt 0 ]]; then
   for jnl in "${Journals[@]}"; do
      Cmd="$P4DBIN -r $P4ROOT -f"
      if [[ "$ReplicaTypeTag" == "edge" ]]; then
         # shellcheck disable=SC2154
         Cmd+=" -K $ExcludedTables"
      fi
      Cmd+=" -jr $jnl"
      msg "Replaying journal to P4ROOT with this command:\\n$Cmd"
      $Cmd || bail "Replay of journal to P4ROOT failed for journal: $jnl"
   done
fi

rm -f "$P4ROOT/P4ROOT_not_usable.txt"

Cmd="$P4DBIN -r $P4ROOT -t localhost:0.0.0.0 -xu"
msg "Ensuring databases are upgraded with this command:\\n$Cmd"
$Cmd || bail "Database upgrade in P4ROOT failed."

if [[ "$LoadOfflineDB" -eq 1 ]]; then
   OfflineDB=${P4ROOT/root/offline_db}
   OfflineDBUsableFile="$OfflineDB/offline_db_usable.txt"

   cd "$OfflineDB" || bail "Could not cd to offline_db dir [$OfflineDB]."

   [[ "$PWD" == "/p4/${SDPInstance}/offline_db" ]] || bail "Unexpected OfflineDB value of: $OfflineDB"

   # Remove db.* and state* files, but not server.id or license files.
   Cmd="/bin/rm -f db.* state* $OfflineDBUsableFile"
   msg "Cleaning up databases and offline_db_usable.txt file with this command in $PWD:\\n$Cmd"
   $Cmd || bail "Failed to clean up old databases and state files in $OfflineDB"

   Cmd="$P4DBIN -r $OfflineDB $CompressFlag $CaseFlag"
   # For edge servers, ignore tables specified by $ExcludedTables, defined in
   # /p4/common/bin/edge_vars.
   if [[ "$ReplicaTypeTag" == "edge" ]]; then
      # shellcheck disable=SC2154
      Cmd+=" -K $ExcludedTables"
   fi
   Cmd+=" -jr $Checkpoint"

   msg "Replaying checkpoint to offline_db with this command:\\n$Cmd"
   $Cmd || bail "Checkpoint replay to offline_db failed."

   Cmd="$P4DBIN -r $OfflineDB -t localhost:0.0.0.0 -xu"
   msg "Ensuring databases in offline_db are upgraded with this command:\\n$Cmd"
   $Cmd || bail "Database upgrade in offline_db failed."

   echo "Database restored successfully." > "$OfflineDBUsableFile" ||\
      bail "Failed to write this file: $OfflineDBUsableFile"
fi

msg "Starting p4d."
if [[ -r "$P4DSystemdServiceFile" ]]; then
   sudo systemctl start "${P4DBIN##*/}" ||\
      bail "Failed to execute: sudo systemctl start ${P4DBIN##*/}"
else
   "$P4DInitScript" start
fi

# Delay a bit to give p4d time to start. As it comes up,
# ensure the local SSL connection is trusted.
i=0; while [[ "$i" -lt 10 ]]; do
   sleep 2
   if p4 info -s > /dev/null 2>&1; then
      break
   elif [[ "$P4PORT" =~ ^ssl[46]*: ]]; then
      p4 trust -f -y > /dev/null 2>&1
   fi
   i+=1
done

if p4 info -s > /dev/null 2>&1; then
   msg "The P4D Server started OK."
else
   msg "Warning: The P4D Server did not start."
   OverallReturnStatus=1
fi

msg "Logging in."
msg "Running: $P4CBIN/p4login -v"
"$P4CBIN/p4login" -v
msg "Running: $P4CBIN/p4login -p $P4MASTERPORT -v"
"$P4CBIN/p4login" -p "$P4MASTERPORT" -v

# Now that we have loaded a checkpoint, re-load the shell environment to
# ensure that P4REPLICA is set correctly, as some db.config settings are
# determined dynamically.

# shellcheck disable=SC1091
source /p4/common/bin/p4_vars "$SDPInstance"
if [[ "$P4REPLICA" == TRUE ]]; then
   msg "Logging in service user."
   msg "Running: $P4CBIN/p4login -v -service"
   "$P4CBIN/p4login" -v -service

   sleep 3
   Cmd="$P4BIN -s pull -lj"
   msg "Checking replication status with this command:\\n$Cmd"
   $Cmd || bail "Replica is not replicating properly."
fi

msg "\\nAll processing completed successfully.\\n"

if [[ "$DoVerify" -eq 1 ]]; then
   Cmd="p4verify.sh $SDPInstance $VerifyOptions"
   msg "Performing archive verification with this command, launched in as a\\nbackground process (fire and forget):\\nnohup $Cmd < /dev/null > /dev/null 2>&1 &"
   # shellcheck disable=SC2086
   nohup $Cmd < /dev/null > /dev/null 2>&1 &

   sleep 1
   msg "\\nThe p4verify.sh log is: $LOGS/p4verify.log"
fi

if [[ -x "$P4BrokerInitScript" && -r "$P4BrokerCfg" ]]; then
   if [[ "$StartBrokerWhenDone" -eq 1 ]]; then
      if [[ -r "$P4BrokerSystemdServiceFile" ]]; then
         sudo systemctl start "${P4BROKERBIN##*/}" ||\
            bail "Failed to execute: sudo systemctl start ${P4BROKERBIN##*/}"
      else
         "$P4BrokerInitScript" start
      fi

      i=0; while [[ "$i" -lt 10 ]]; do
         sleep 3
         msg "Running (try $((i+1))): $P4CBIN/p4login -p $P4BROKERPORT -v"
         "$P4CBIN/p4login" -p "$P4BROKERPORT" -v && break
         i+=1
      done
   else
      if [[ -r "$P4BrokerSystemdServiceFile" ]]; then
         msg "NOT starting broker. Start it manually when ready with:\\n\\tsudo systemctl start ${P4BROKERBIN##*/}\\n"
      else
         msg "NOT starting broker. Start it manually when ready with:\\n\\t$P4BrokerInitScript start\\n"
      fi
   fi
fi

msg "Checkpoint load processing took $((SECONDS/3600)) hours $((SECONDS%3600/60)) minutes $((SECONDS%60)) seconds.\\n"

exit "$OverallReturnStatus"