--- /dev/null
+ Technical notes on version 1.7 10Dec01
+ 12 December 2001
+ Kern Sibbald
+
+General:
+- The size keeps growing: Total files = 169 Total lines = 40504
+- All daemons must be updated at the same time as the
+ Job Start protocol between the Director and File daemon
+ changed.
+- Major rework of how JCRs are handled so that they
+ are always locked when walking the JCR chain.
+- JCRs now have use_count, and global free_jcr() routine,
+ with daemon specific routine provided in new_jcr() call.
+- Storage daemon is now Job reentrant.
+- Updated the Copyright to include 2002.
+- Tested RESTORE, fixed to handle DOS restore to Unix,
+ and fixed some directory permission creation bugs.
+
+Changes submitted this submission:
+- Implemented a general db_list_sql_query() so that
+ the Console program can submit arbitrary catalog
+ queries.
+- Reworked rl_catch_signals so that it works on Win32.
+- Added authentication between the Storage daemon and the
+ Client. Also provide a random authorization key.
+- Consolidated more variables in the global part of the
+ JCR record.
+- Improved setting JobStatus.
+- Storage daemon now passes JobStatus changes back to the
+ Director.
+- Removed obsoleted File daemon job id from Job Start protocol.
+- Require confirm on Cancel command if there is only
+ one Job to be cancelled.
+- Added a prompt for the run command when running a restore
+ job.
+- Added two new Console commands:
+
+ sqlquery prompts for a command and submits it
+ to the SQL engine.
+
+ query prompts for submitting a "canned" command.
+ There are currently 3 canned commands.
+
+ "List where a file is saved"
+ "List total files/bytes by Job"
+ "List total files/bytes by Volume"
+
+- Improved Status output -- more precise information on wait
+ conditions when doing Director status. More improvements
+ to come.
+- File deamon now reports "Files examined" in Status output.
+ This is significant when doing an incremental backup where
+ a lot of files are examined but few are saved.
+- Fixed a bad status check when opening a drive in the
+ Storage daemon (forgot to change for updated calling sequence).
+- When RESTORING if a file could not be created, there were a
+ pile of error messages due to a bug with the stream not being
+ synchronized -- fixed.
+- Removed debug code that set the interval between mount messages
+ to 1 hour -- this bit John.