Index: TODO ================================================================== --- TODO +++ TODO @@ -22,14 +22,14 @@ . Use file semaphore to kill tests, eliminate db load of the KILLREQ query . Merge this change to revolution branch 23WW45 - the revolution branch . Add "fast" db start option (no handshaking over NFS) . Add server-ro to server types (just "server" is fine for read/write). -. Create pause-server and resume-server calls +. [DONE] Create pause-server and resume-server calls . Create rsync or cp sync to MTRAH function . Change rmt:send-receive to divert calls to read-only server when possible -. Change start server to call main.db server for 1..N.db servers, block until server is read for use. +. [DONE] Change start server to call main.db server for 1..N.db servers, block until server is read for use. 23WW21 . Dashboard needs its own cache db in /tmp 23WW07 Index: dbmod.scm ================================================================== --- dbmod.scm +++ dbmod.scm @@ -88,13 +88,13 @@ ;; The cachedb one-db file per server method goes in here ;;====================================================================== ;; NOTE: the r/w is now w/r, #t=db modified by query, #f=db NOT modified by query (define (dbmod:with-db dbstruct run-id w/r proc params) - (let* ((use-mutex (or (and w/r ;; use the mutex on queries that modify the db and for sync to disk - (> *api-process-request-count* 5)) ;; when writes are happening throttle more - (> *api-process-request-count* 50))) + (let* ((use-mutex w/r) ;; (or (and w/r ;; use the mutex on queries that modify the db and for sync to disk + ;; (> *api-process-request-count* 5)) ;; when writes are happening throttle more + ;; (> *api-process-request-count* 50))) (dbdat (dbmod:open-db dbstruct run-id (dbfile:db-init-proc))) (dbh (dbr:dbdat-dbh dbdat)) ;; this will be the cachedb handle (dbfile (dbr:dbdat-dbfile dbdat))) ;; if nfs mode do a sync if delta > 2 (let* ((last-update (dbr:dbstruct-last-update dbstruct))