Index: Makefile
==================================================================
--- Makefile
+++ Makefile
@@ -163,10 +163,15 @@
 	chmod a+x $@
 
 deploytarg/nbfind : utils/nbfind
 	$(INSTALL) $< $@
 	chmod a+x $@
+
+$(PREFIX)/bin/mtest-reaper: helpers/mtest-reaper.scm helpers/ducttape-lib.scm helpers/inteldate.scm helpers/mimetypes.scm
+	make -C helpers $@ PREFIX=$(PREFIX) INSTALL=$(INSTALL) ARCHSTR=$(ARCHSTR)
+
+mtest-reaper: $(PREFIX)/bin/mtest-reaper
 
 # install dashboard as dboard so wrapper script can be called dashboard
 $(PREFIX)/bin/.$(ARCHSTR)/dboard : dboard $(FILES) utils/mk_wrapper
 	utils/mk_wrapper $(PREFIX) dboard $(PREFIX)/bin/dashboard
 	chmod a+x $(PREFIX)/bin/dashboard
@@ -177,10 +182,11 @@
 	  $(PREFIX)/bin/nbfind $(PREFIX)/bin/loadrunner $(PREFIX)/bin/viewscreen $(PREFIX)/bin/mt_xterm \
 	  $(PREFIX)/share/docs/megatest_manual.html $(PREFIX)/bin/remrun
 
 $(PREFIX)/bin/.$(ARCHSTR) : 
 	mkdir -p $(PREFIX)/bin/.$(ARCHSTR)
+	mkdir -p $(PREFIX)/bin/.$(ARCHSTR)/lib
 
 test: tests/tests.scm
 	cd tests;csi -I .. -b -n tests.scm
 
 ext-tests/.fslckout : $(MTQA_FOSSIL)
@@ -282,5 +288,6 @@
 	   echo "(use postgresql)(hash-table-set! *available-db* 'postgresql #t)" >> altdb.scm;\
 	fi
 
 portlogger-example : portlogger-example.scm api.o archive.o client.o common.o configf.o daemon.o dashboard-tests.o db.o dcommon.o ezsteps.o filedb.o genexample.o gutils.o http-transport.o items.o keys.o launch.o lock-queue.o margs.o megatest-version.o mt.o ods.o portlogger.o process.o rmt.o rpc-transport.o runconfig.o runs.o sdb.o server.o synchash.o tasks.o tdb.o tests.o tree.o
 	csc $(CSCOPTS) portlogger-example.scm api.o archive.o client.o common.o configf.o daemon.o dashboard-tests.o db.o dcommon.o ezsteps.o filedb.o genexample.o gutils.o http-transport.o items.o keys.o launch.o lock-queue.o margs.o megatest-version.o mt.o ods.o portlogger.o process.o rmt.o rpc-transport.o runconfig.o runs.o sdb.o server.o synchash.o tasks.o tdb.o tests.o tree.o
+

Index: api.scm
==================================================================
--- api.scm
+++ api.scm
@@ -62,10 +62,12 @@
     synchash-get
     ))
 
 (define api:write-queries
   '(
+    get-keys-write ;; dummy "write" query to force server start
+
     ;; SERVERS
     start-server
     kill-server
 
     ;; TESTS
@@ -189,10 +191,11 @@
 	    ;;======================================================================
 
 	    ;; KEYS
 	    ((get-key-val-pairs)               (apply db:get-key-val-pairs dbstruct params))
 	    ((get-keys)                        (db:get-keys dbstruct))
+            ((get-keys-write)                        (db:get-keys dbstruct)) ;; force a dummy "write" query to force server
 	    ((get-key-vals)                    (apply db:get-key-vals dbstruct params))
 	    ((get-target)                      (apply db:get-target dbstruct params))
 	    ((get-targets)                     (db:get-targets dbstruct))
 
 	    ;; ARCHIVES

Index: common.scm
==================================================================
--- common.scm
+++ common.scm
@@ -103,14 +103,13 @@
 (define *db-stats-mutex*      (make-mutex))
 ;; db access
 (define *db-last-access*      (current-seconds)) ;; last db access, used in server
 (define *db-write-access*     #t)
 ;; db sync
-(define *db-last-write*       0)                 ;; used to record last touch of db
 (define *db-last-sync*        0)                 ;; last time the sync to megatest.db happened
 (define *db-sync-in-progress* #f)                ;; if there is a sync in progress do not try to start another
-(define *db-multi-sync-mutex* (make-mutex))      ;; protect access to *db-sync-in-progress*, *db-last-sync* and *db-last-write*
+(define *db-multi-sync-mutex* (make-mutex))      ;; protect access to *db-sync-in-progress*, *db-last-sync*
 ;; task db
 (define *task-db*             #f) ;; (vector db path-to-db)
 (define *db-access-allowed*   #t) ;; flag to allow access
 (define *db-access-mutex*     (make-mutex))
 (define *db-cache-path*       #f)
@@ -159,10 +158,18 @@
 
 (define *run-info-cache*     (make-hash-table)) ;; run info is stable, no need to reget
 (define *launch-setup-mutex* (make-mutex))     ;; need to be able to call launch:setup often so mutex it and re-call the real deal only if *toppath* not set
 (define *homehost-mutex*     (make-mutex))
 
+(defstruct remote
+  (hh-dat            (common:get-homehost)) ;; homehost record ( addr . hhflag )
+  (server-url        (if *toppath* (server:read-dotserver *toppath*))) ;; (server:check-if-running *toppath*) #f))
+  (last-server-check 0)  ;; last time we checked to see if the server was alive
+  (conndat           #f)
+  (transport         *transport-type*)
+  (server-timeout    (or (server:get-timeout) 100))) ;; default to 100 seconds
+
 ;; launching and hosts
 (defstruct host
   (reachable    #f)
   (last-update  0)
   (last-used    0)
@@ -555,26 +562,28 @@
 ;;======================================================================
 ;; E X I T   H A N D L I N G
 ;;======================================================================
 
 (define (common:run-sync?)
-  (let ((ohh (common:on-homehost?))
-	(srv (args:get-arg "-server")))
-    ;; (debug:print-info 0 *default-log-port* "common:run-sync? ohh=" ohh ", srv=" srv)
     (and (common:on-homehost?)
-	 (args:get-arg "-server"))))
+	 (args:get-arg "-server")))
+
+;;   (let ((ohh (common:on-homehost?))
+;; 	(srv (args:get-arg "-server")))
+;;     (and ohh srv)))
+    ;; (debug:print-info 0 *default-log-port* "common:run-sync? ohh=" ohh ", srv=" srv)
 
 ;;;; run-ids
 ;;    if #f use *db-local-sync* : or 'local-sync-flags
 ;;    if #t use timestamps      : or 'timestamps
 (define (common:sync-to-megatest.db dbstruct) 
   (let ((start-time         (current-seconds))
 	(res                (db:multi-db-sync dbstruct 'new2old)))
     (let ((sync-time (- (current-seconds) start-time)))
-      (debug:print-info 3 *default-log-port* "Sync of newdb to olddb completed in " sync-time " seconds")
+      (debug:print-info 3 *default-log-port* "Sync of newdb to olddb completed in " sync-time " seconds pid="(current-process-id))
       (if (common:low-noise-print 30 "sync new to old")
-	  (debug:print-info 0 *default-log-port* "Sync of newdb to olddb completed in " sync-time " seconds")))
+	  (debug:print-info 0 *default-log-port* "Sync of newdb to olddb completed in " sync-time " seconds pid="(current-process-id))))
     res))
 
 ;; currently the primary job of the watchdog is to run the sync back to megatest.db from the db in /tmp
 ;; if we are on the homehost and we are a server (by definition we are on the homehost if we are a server)
 ;;
@@ -581,19 +590,20 @@
 (define (common:watchdog)
   (thread-sleep! 0.05) ;; delay for startup
   (let ((legacy-sync (common:run-sync?))
 	(debug-mode  (debug:debug-mode 1))
 	(last-time   (current-seconds)))
-    (debug:print-info 0 *default-log-port* "watchdog starting. legacy-sync is " legacy-sync)
+    (debug:print-info 0 *default-log-port* "watchdog starting. legacy-sync is " legacy-sync" pid="(current-process-id))
     (if legacy-sync
 	(let ((dbstruct (db:setup)))
 	  (debug:print-info 0 *default-log-port* "Server running, periodic sync started.")
 	  (let loop ()
+            ;;(BB> "watchdog loop.  pid="(current-process-id))
 	    ;; sync for filesystem local db writes
 	    ;;
 	    (mutex-lock! *db-multi-sync-mutex*)
-	    (let* ((need-sync        (>= *db-last-write* *db-last-sync*)) ;; no sync since last write
+	    (let* ((need-sync        (>= *db-last-access* *db-last-sync*)) ;; no sync since last write
 		   (sync-in-progress *db-sync-in-progress*)
 		   (should-sync      (> (- (current-seconds) *db-last-sync*) 5)) ;; sync every five seconds minimum
 		   (will-sync        (and (or need-sync should-sync)
 					  (not sync-in-progress)))
 		   (start-time       (current-seconds)))
@@ -630,41 +640,48 @@
 		      (begin
 			(thread-sleep! 1)
 			(delay-loop (+ count 1))))
 		  (loop)))
 	    (if (common:low-noise-print 30)
-		(debug:print-info 0 *default-log-port* "Exiting watchdog timer, *time-to-exit* = " *time-to-exit*)))))))
+		(debug:print-info 0 *default-log-port* "Exiting watchdog timer, *time-to-exit* = " *time-to-exit*" pid="(current-process-id))))))))
 
 (define (std-exit-procedure)
-  
+  (on-exit (lambda () 0))
   (let ((no-hurry  (if *time-to-exit* ;; hurry up
 		       #f
 		       (begin
 			 (set! *time-to-exit* #t)
 			 #t))))
     (debug:print-info 4 *default-log-port* "starting exit process, finalizing databases.")
     (if (and no-hurry (debug:debug-mode 18))
 	(rmt:print-db-stats))
     (let ((th1 (make-thread (lambda () ;; thread for cleaning up, give it five seconds
-			      (if *dbstruct-db* (db:close-all *dbstruct-db*)) ;; one second allocated
+                              (if *dbstruct-db* (db:close-all *dbstruct-db*)) ;; one second allocated
 			      (if *task-db*    
 				  (let ((db (cdr *task-db*)))
 				    (if (sqlite3:database? db)
 					(begin
 					  (sqlite3:interrupt! db)
 					  (sqlite3:finalize! db #t)
 					  ;; (vector-set! *task-db* 0 #f)
 					  (set! *task-db* #f)))))
-			      (close-output-port *default-log-port*)
+                              (if (and *runremote*
+                                       (remote-conndat *runremote*))
+                                  (begin
+                                    (http-client#close-all-connections!))) ;; for http-client
+                              (if (not (eq? *default-log-port* (current-error-port)))
+                                  (close-output-port *default-log-port*))
 			      (set! *default-log-port* (current-error-port))) "Cleanup db exit thread"))
 	  (th2 (make-thread (lambda ()
 			      (debug:print 4 *default-log-port* "Attempting clean exit. Please be patient and wait a few seconds...")
 			      (if no-hurry
-				  (thread-sleep! 5) ;; give the clean up few seconds to do it's stuff
-				  (thread-sleep! 2))
-			      (debug:print 4 *default-log-port* " ... done")
-			      )
+                                  (begin
+                                    (thread-sleep! 5)) ;; give the clean up few seconds to do it's stuff
+                                  (begin
+      				  (thread-sleep! 2)))
+      			      (debug:print 4 *default-log-port* " ... done")
+      			      )
 			    "clean exit")))
 
       ;; let's try to clean up open sockets
       (if *runremote*
           (case (remote-transport *runremote*)
@@ -673,11 +690,15 @@
             (else
              (debug:print-info 0 *default-log-port* "Transport "(remote-transport *runremote*)" not supported"))))
 
       (thread-start! th1)
       (thread-start! th2)
-      (thread-join! th1))))
+      (thread-join! th1)
+      )
+    )
+
+  0)
 
 (define (std-signal-handler signum)
   ;; (signal-mask! signum)
   (set! *time-to-exit* #t)
   (debug:print-error 0 *default-log-port* "Received signal " signum " exiting promptly")
@@ -813,20 +834,24 @@
 
 (define (common:args-get-status)
   (or (args:get-arg "-status")(args:get-arg ":status")))
 
 (define (common:args-get-testpatt rconf)
-  (let* ((rtestpatt     (if rconf (runconfigs-get rconf "TESTPATT") #f))
-	 (args-testpatt (or (args:get-arg "-testpatt")
-			    (args:get-arg "-runtests")
-			    "%"))
-	 (testpatt    (or (and (equal? args-testpatt "%")
-			       rtestpatt)
-			  args-testpatt)))
-    (if rtestpatt (debug:print-info 0 *default-log-port* "TESTPATT from runconfigs: " rtestpatt))
-    testpatt))
-
+  (let* ((tagexpr (args:get-arg "-tagexpr"))
+         (tags-testpatt (if tagexpr (string-join (runs:get-tests-matching-tags tagexpr) ",") #f))
+         (testpatt-key  (if (args:get-arg "-mode") (args:get-arg "-mode") "TESTPATT"))
+         (args-testpatt (or (args:get-arg "-testpatt") (args:get-arg "-runtests") "%"))
+         (rtestpatt     (if rconf (runconfigs-get rconf testpatt-key) #f)))
+    (cond
+     (tags-testpatt
+      (debug:print-info 0 *default-log-port* "-tagexpr "tagexpr" selects testpatt "tags-testpatt)
+      tags-testpatt)
+     ((and (equal? args-testpatt "%") rtestpatt)
+      (debug:print-info 0 *default-log-port* "testpatt defined in "testpatt-key" from runconfigs: " rtestpatt)
+      rtestpatt)
+     (else args-testpatt))))
+     
 (define (common:get-linktree)
   (or (getenv "MT_LINKTREE")
       (if *configdat*
 	  (configf:lookup *configdat* "setup" "linktree"))))
 

Index: db.scm
==================================================================
--- db.scm
+++ db.scm
@@ -335,10 +335,11 @@
     (let ((update_info (cons (if force-sync 0 *db-last-sync*) "last_update")))
       (mutex-unlock! *db-multi-sync-mutex*)
       (db:sync-tables (db:sync-all-tables-list dbstruct) update_info tmpdb refndb mtdb))
     (mutex-lock! *db-multi-sync-mutex*)
     (set! *db-last-sync* start-t)
+    (set! *db-last-access* start-t)
     (mutex-unlock! *db-multi-sync-mutex*)))
 
 ;; close all opened run-id dbs
 (define (db:close-all dbstruct)
   (if (dbr:dbstruct? dbstruct)

Index: http-transport.scm
==================================================================
--- http-transport.scm
+++ http-transport.scm
@@ -486,11 +486,11 @@
 ;; 	(thread-sleep! rem-time)
 ;; 	(thread-sleep! 4))) ;; fallback for if the math is changed ...
 
 (define (http-transport:server-shutdown server-id port)
   (let ((tdbdat (tasks:open-db)))
-    (debug:print-info 0 *default-log-port* "Starting to shutdown the server.")
+    (debug:print-info 0 *default-log-port* "Starting to shutdown the server. pid="(current-process-id))
     ;;
     ;; start_shutdown
     ;;
     (tasks:server-set-state! (db:delay-if-busy tdbdat) server-id "shutting-down")
     (set! *time-to-exit* #t) ;; tell on-exit to be fast as we've already cleaned up

DELETED inteldate.scm
Index: inteldate.scm
==================================================================
--- inteldate.scm
+++ /dev/null
@@ -1,180 +0,0 @@
-(use srfi-19)
-(use test)
-(use format)
-(use regex)
-(declare (unit inteldate))
-;; utility procedures to convert among
-;; different ways to express date (inteldate, seconds since epoch, isodate)
-;;
-;; samples:
-;; isodate   -> "2016-01-01"
-;; inteldate -> "16ww01.5"
-;; seconds   -> 1451631600
-
-;; procedures provided:
-;; ====================
-;; seconds->isodate
-;; seconds->inteldate
-;;
-;; isodate->seconds
-;; isodate->inteldate
-;;
-;; inteldate->seconds
-;; inteldate->isodate
-
-;; srfi-19 used extensively; this doc is better tha the eggref:
-;; http://srfi.schemers.org/srfi-19/srfi-19.html
-
-;; Author: brandon.j.barclay@intel.com 16ww18.6
-
-(define (date->seconds date)
-  (inexact->exact
-   (string->number
-    (date->string date "~s"))))
-
-(define (seconds->isodate seconds)
-  (let* ((date (seconds->date seconds))
-         (result (date->string date "~Y-~m-~d")))
-    result))
-
-(define (isodate->seconds isodate)
-  "Takes a string input of the form 'YY-MM-DD' or 'YYYY-MM-DD' and returns epoch time; for YY, assume after Y2K"
-  (let* ((numlist (map string->number (string-split isodate "-")))
-        (raw-year (car numlist))
-        (year (if (< raw-year 100) (+ raw-year 2000) raw-year))
-        (month (list-ref numlist 1))
-        (day (list-ref numlist 2))
-        (date (make-date 0 0 0 0 day month year))
-        (seconds (date->seconds date)))
-
-    seconds))
-
-;; adapted from perl Intel::WorkWeek perl module
-;; intel year consists of numbered weeks starting from week 1
-;;   week 1 is the week containing jan 1 of the year
-;;   days of week are numbered starting from 0 on sunday
-;;   intel year does not match calendar year in workweek 1
-;;     before jan1.
-(define (seconds->inteldate-values seconds)
-  (define (date-difference->seconds d1 d2)
-    (- (date->seconds d1) (date->seconds d2)))
-
-  (let* ((thisdate (seconds->date seconds))
-         (thisdow (string->number (date->string thisdate "~w")))
-
-         (year (date-year thisdate))
-         ;; intel workweek 1 begins on sunday of week containing jan1
-         (jan1 (make-date 0 0 0 0 1 1 year))
-         (jan1dow (date-week-day jan1))
-         (ww01 (date-subtract-duration jan1 (seconds->time (* 60 60 24 jan1dow))))
-
-         (ww01_delta_seconds (date-difference->seconds thisdate ww01))
-         (wwnum_initial (inexact->exact (add1 (floor (/ ww01_delta_seconds 24 3600 7) ))))
-         
-         ;; we could be in ww1 of next year
-         (this-saturday (seconds->date
-                         (+ seconds
-                            (* 60 60 24 (- 6 thisdow)))))
-         (this-week-ends-next-year?
-          (> (date-year this-saturday) year))
-         (intelyear
-          (if this-week-ends-next-year?
-              (add1 year)
-              year))
-         (intelweek
-          (if this-week-ends-next-year?
-              1
-              wwnum_initial)))
-   (values intelyear intelweek thisdow)))
-
-(define (seconds->inteldate seconds)
-  (define (string-leftpad in width pad-char)
-    (let* ((unpadded-str (->string in))
-           (padlen_temp (- width (string-length unpadded-str)))
-           (padlen (if (< padlen_temp 0) 0 padlen_temp))
-           (padding
-            (fold conc ""
-                  (map (lambda (x) (->string pad-char)) (iota padlen)))))
-      (conc padding unpadded-str)))
-  (define (zeropad num width)
-    (string-leftpad num width #:0))
-
-  (let-values (((intelyear intelweek day-of-week-num)
-                (seconds->inteldate-values seconds)))
-    (let ((intelyear-str
-           (zeropad
-            (->string
-             (if (> intelyear 1999)
-                 (- intelyear 2000) intelyear))
-            2))
-          (intelweek-str
-           (zeropad (->string intelweek) 2))
-          (dow-str (->string day-of-week-num)))
-      (conc intelyear-str "ww" intelweek-str "." dow-str))))
-
-(define (isodate->inteldate isodate)
-  (seconds->inteldate
-   (isodate->seconds isodate)))
-
-(define (inteldate->seconds inteldate)
-  (let ((match (string-match "^(\\d+)ww(\\d+).(\\d)$" inteldate)))
-    (if
-     (not match)
-     #f
-     (let* (
-            (intelyear-raw (string->number (list-ref match 1)))
-            (intelyear (if (< intelyear-raw 100)
-                           (+ intelyear-raw 2000)
-                           intelyear-raw))
-            (intelww (string->number (list-ref match 2)))
-            (dayofweek (string->number (list-ref match 3)))
-
-            (day-of-seconds (* 60 60 24 ))
-            (week-of-seconds (* day-of-seconds 7))
-            
-
-            ;; get seconds at ww1.0
-            (new-years-date (make-date 0 0 0 0 1 1 intelyear))
-            (new-years-seconds
-             (date->seconds new-years-date))
-            (new-years-dayofweek (date-week-day new-years-date))
-            (ww1.0_seconds (- new-years-seconds
-                              (* day-of-seconds
-                                 new-years-dayofweek)))
-            (workweek-adjustment (* week-of-seconds (sub1 intelww)))
-            (weekday-adjustment (* dayofweek day-of-seconds))
-
-            (result (+ ww1.0_seconds workweek-adjustment weekday-adjustment)))
-       result))))
-
-(define (inteldate->isodate inteldate)
-  (seconds->isodate (inteldate->seconds inteldate)))
-
-(define (inteldate-tests)
-  (test-group
-   "date conversion tests"
-   (let ((test-table
-          '(("16ww01.5" . "2016-01-01")
-            ("16ww18.5" . "2016-04-29")
-            ("1999ww33.5" . "1999-08-13")
-            ("16ww18.4" . "2016-04-28")
-            ("16ww18.3" . "2016-04-27")
-            ("13ww01.0" . "2012-12-30")
-            ("13ww52.6" . "2013-12-28")
-            ("16ww53.3" . "2016-12-28"))))
-     (for-each
-      (lambda (test-pair)
-        (let ((inteldate (car test-pair))
-              (isodate (cdr test-pair)))
-          (test
-           (conc "(isodate->inteldate "isodate ") => "inteldate)
-           inteldate
-           (isodate->inteldate isodate))
-          
-          (test
-           (conc "(inteldate->isodate "inteldate ")   => "isodate)
-           isodate
-           (inteldate->isodate inteldate))))
-      test-table))))
-
-;(inteldate-tests)

Index: launch.scm
==================================================================
--- launch.scm
+++ launch.scm
@@ -859,11 +859,14 @@
 	     (directory-exists? *toppath*))
 	(begin
 	  (setenv "MT_RUN_AREA_HOME" *toppath*)
 	  (setenv "MT_TESTSUITE_NAME" (common:get-testsuite-name)))
 	(begin
-	  (debug:print-error 0 *default-log-port* "failed to find the top path to your Megatest area.")))
+	  (debug:print-error 0 *default-log-port* "failed to find the top path to your Megatest area.")
+          ;;(exit 1)
+          #f
+          ))
     *toppath*))
 
 (define (get-best-disk confdat testconfig)
   (let* ((disks   (or (and testconfig (hash-table-ref/default testconfig "disks" #f))
 		      (hash-table-ref/default confdat "disks" #f)))

Index: megatest.scm
==================================================================
--- megatest.scm
+++ megatest.scm
@@ -93,10 +93,12 @@
   -reqtarg key1/key2/...  : run for key1, key2, etc. but key1/key2 must be in runconfig
   -testpatt patt1/patt2,patt3/...  : % is wildcard
   -runname                : required, name for this particular test run
   -state                  : Applies to runs, tests or steps depending on context
   -status                 : Applies to runs, tests or steps depending on context
+  -mode key               : load testpatt from <key> in runconfigs instead of default TESTPATT
+  -tagexpr tag1,tag2%,..  : select tests with tags matching expression
 
 Test helpers (for use inside tests)
   -step stepname
   -test-status            : set the state and status of a test (use :state and :status)
   -setlog logfname        : set the path/filename to the final log relative to the test
@@ -209,11 +211,13 @@
 			":state"  
 			"-state"
 			":status"
 			"-status"
 			"-list-runs"
-			"-testpatt" 
+			"-testpatt"
+                        "-mode"
+                        "-tagexpr"
 			"-itempatt"
 			"-setlog"
 			"-set-toplog"
 			"-runstep"
 			"-logpro"
@@ -1983,16 +1987,15 @@
 
 ;;======================================================================
 ;; Exit and clean up
 ;;======================================================================
 
-
 (if (not *didsomething*)
     (debug:print 0 *default-log-port* help))
 
-(set! *time-to-exit* #t)
 (thread-join! *watchdog*)
+(set! *time-to-exit* #t)
 
 (if (not (eq? *globalexitstatus* 0))
     (if (or (args:get-arg "-run")(args:get-arg "-runtests")(args:get-arg "-runall"))
         (begin
            (debug:print 0 *default-log-port* "NOTE: Subprocesses with non-zero exit code detected: " *globalexitstatus*)

Index: rmt.scm
==================================================================
--- rmt.scm
+++ rmt.scm
@@ -25,18 +25,10 @@
 
 ;; generate entries for ~/.megatestrc with the following
 ;;
 ;;  grep define ../rmt.scm | grep rmt: |perl -pi -e 's/\(define\s+\((\S+)\W.*$/\1/'|sort -u
 
-(defstruct remote
-  (hh-dat            (common:get-homehost)) ;; homehost record ( addr . hhflag )
-  (server-url        (if *toppath* (server:read-dotserver *toppath*))) ;; (server:check-if-running *toppath*) #f))
-  (last-server-check 0)  ;; last time we checked to see if the server was alive
-  (conndat           #f)
-  (transport         *transport-type*)
-  (server-timeout    (or (server:get-timeout) 100))) ;; default to 100 seconds
-
 ;;======================================================================
 ;;  S U P P O R T   F U N C T I O N S
 ;;======================================================================
 
 ;; if a server is either running or in the process of starting call client:setup
@@ -95,23 +87,38 @@
      ((and (cdr (remote-hh-dat *runremote*))   ;; on homehost
            (member cmd api:read-only-queries)) ;; this is a read
       (mutex-unlock! *rmt-mutex*)
       (debug:print-info 12 *default-log-port* "rmt:send-receive, case  3")
       (rmt:open-qry-close-locally cmd 0 params))
+
+     ;; on homehost and this is a write, we already have a server, but server has died
+     ((and (cdr (remote-hh-dat *runremote*))         ;; on homehost
+           (not (member cmd api:read-only-queries))  ;; this is a write
+           (remote-server-url *runremote*)           ;; have a server
+           (not (server:read-dotserver *toppath*)))  ;; server has died.
+      (set! *runremote* #f)
+      (mutex-unlock! *rmt-mutex*)
+      (debug:print-info 12 *default-log-port* "rmt:send-receive, case  4.1")
+      (rmt:send-receive cmd rid params attemptnum: attemptnum))
+
      ;; on homehost and this is a write, we already have a server
      ((and (cdr (remote-hh-dat *runremote*))         ;; on homehost
            (not (member cmd api:read-only-queries))  ;; this is a write
            (remote-server-url *runremote*))          ;; have a server
       (mutex-unlock! *rmt-mutex*)
       (debug:print-info 12 *default-log-port* "rmt:send-receive, case  4")
       (rmt:open-qry-close-locally cmd 0 params))
-     ;; on homehost and this is a write, we have a server (we know because case 4 checked)
-     ((and (cdr (remote-hh-dat *runremote*))         ;; on homehost
-	   (not (member cmd api:read-only-queries)))
-      (mutex-unlock! *rmt-mutex*)
-      (debug:print-info 12 *default-log-port* "rmt:send-receive, case  4.1")
-      (rmt:open-qry-close-locally cmd 0 params))
+
+     ;; commented by bb; this was blocking server passive start on write on homehost (case 5)
+     ;; ;; on homehost and this is a write, we have a server (we know because case 4 checked)
+     ;; ((and (cdr (remote-hh-dat *runremote*))         ;; on homehost
+     ;;       (not (member cmd api:read-only-queries)))
+     ;;  (mutex-unlock! *rmt-mutex*)
+     ;;  (debug:print-info 12 *default-log-port* "rmt:send-receive, case  4.1")
+     ;;  (rmt:open-qry-close-locally cmd 0 params))
+
+     
      ;; no server contact made and this is a write, passively start a server 
      ((and (not (remote-server-url *runremote*))
 	   (not (member cmd api:read-only-queries)))
       (debug:print-info 12 *default-log-port* "rmt:send-receive, case  5")
       (let ((serverconn (server:read-dotserver *toppath*))) ;; (server:check-if-running *toppath*))) ;; Do NOT want to run server:check-if-running - very expensive to do for every write call
@@ -267,11 +274,11 @@
 	  ;; (rmt:update-db-stats run-id cmd params duration)
 	  ;; mark this run as dirty if this was a write, the watchdog is responsible for syncing it
 	  (if qry-is-write
 	      (let ((start-time (current-seconds)))
 		(mutex-lock! *db-multi-sync-mutex*)
-		(set! *db-last-write* start-time) ;; the oldest "write"
+		(set! *db-last-access* start-time)  ;; THIS IS PROBABLY USELESS? (we are on a client)
                 (mutex-unlock! *db-multi-sync-mutex*)))))
     res))
 
 (define (rmt:send-receive-no-auto-client-setup connection-info cmd run-id params)
   (let* ((run-id   (if run-id run-id 0))
@@ -380,10 +387,15 @@
 (define (rmt:get-keys)
   (if *db-keys* *db-keys* 
      (let ((res (rmt:send-receive 'get-keys #f '())))
        (set! *db-keys* res)
        res)))
+
+(define (rmt:get-keys-write) ;; dummy query to force server start
+  (let ((res (rmt:send-receive 'get-keys-write #f '())))
+    (set! *db-keys* res)
+    res))
 
 ;; we don't reuse run-id's (except possibly *after* a db cleanup) so it is safe
 ;; to cache the resuls in a hash
 ;;
 (define (rmt:get-key-vals run-id)