kes Remember what volume is in VTLs (previously turned off for

     testing).
kes  Analysis of JobMedia records by Tom Ivar Helbekkmo <tih@hamartun.priv.no>
     pointed out some problems in creating JobMedia records during
     spooling with concurrent jobs. Fix the problem by updating the
     start file position before despooling and generate a jobmedia
     record at the end of despooling.

git-svn-id: https://bacula.svn.sourceforge.net/svnroot/bacula/trunk@6708 91ce42f0-d328-0410-95d8-f526ca767f89

diff --git a/bacula/kernstodo b/bacula/kernstodo
index e15506656f620ecab7ddded337266b5f68206a4b..f78e655e78947de453c2c56ebbea1ab643904dd6 100644 (file)
--- a/bacula/kernstodo
+++ b/bacula/kernstodo
@@ -70,20 +70,40 @@ Professional Needs:
    and http://www.openeyet.nl/scc/ for managing customer changes
 
 Priority:
+================
+Tom Ivar Helbekkmo <tih@hamartun.priv.no>
 > There's definitely something fishy in the recording of start and
-> end blocks in the JOBMEDIA table.  This is a snip from last night's
-> incremental run (still using 2.2.8 plus the four published patches, plus
-> my posted fix for the jobmedia patch):
+> end blocks in the JOBMEDIA table. 
+> - If several jobs start spooling at the same time, they will all get the
+>   current tape position noted as the StartFile/StartBlock for the job.
+>   If they end up despooling to the file that was current when they
+>   started spooling, this is what will end up in the JOBMEDIA table.  If
+>   there is a file change before they despool, the setting of NewFile in
+>   the dcr structure will fix this up later, but the "start of session"
+>   label is already in the spool file, of course, so it holds the wrong
+>   information anyway.
 >
->  JmId | JID | MId | FI | LI | Sfile | Efile | Sblock | Eblock
->- 119 |   26 |   3 |  1 |  53 |   31 |    31 |      0 |    32 
-   120 |   27 |   3 |  1 |  83 |   31 |    31 |      0 |    242
-   121 |   28 |   3 |  1 | 239 |   31 |    31 |      0 |   5683
+> - If the job is longer than the maximum spool size, it will get its
+>   first spool session despooled, and then start spooling again after the
+>   first despooling is over.  The last blocks despooled to tape from the
+>   first session will not have been recorded, but they will be flushed
+>   later, when the next session despools.  However, if another job has
+>   been despooling while this one is spooling its second round, the
+>   session label written to the spool file at its close will cause the
+>   EndFile/Endblock to be set to wherever the tape is at that time.  When
+>   the dangling record is flushed to JOBMEDIA, it gets this wrong
+>   information.  Both session labels in the spool file will be wrong,
+>   too, of course, because they reflect the state of the tape during
+>   spooling, not during despooling.
 >
-> Since I'm using spooling, those jobs should not be interspersed on
-> tape.  Still, at least it seems the error is in including too many
-> blocks in the set that a job's files occupies, so if I understand
-> correctly, it shouldn't cause any restore problems.  :)
+> I would have to study the code much more closely to work out what's the
+> proper fix -- but it seems clear that it should involve creating the
+> session labels only when something is actually written to the archive
+> device, not during spooling.  I'm tempted to try making do_append_data()
+> not create session labels if we're spooling, and add the making of them
+> to despool_data() in stored/spool.c.  Sound reasonable?
+===========
+
 
 - Re-check new dcr->reserved_volume
 - Softlinks that point to non-existent file are not restored in restore all,

diff --git a/bacula/src/stored/reserve.c b/bacula/src/stored/reserve.c
index f41658d7e54f607d9ce7edade6bf4514f69a583a..881a5a751d0cf86a7667f262ce5727cd5ac5b612 100644 (file)
--- a/bacula/src/stored/reserve.c
+++ b/bacula/src/stored/reserve.c
@@ -540,7 +540,7 @@ bool volume_unused(DCR *dcr)
    dev->vol->released = true;
    Dmsg2(dbglvl, "=== set released. Vol=%s dev=%s\n", dev->vol->vol_name,
          dev->print_name());
-   if (dev->is_tape()) { // || dev->is_autochanger()) {
+   if (dev->is_tape() || dev->is_autochanger()) {
       return true;
    } else {
       /*

diff --git a/bacula/src/stored/spool.c b/bacula/src/stored/spool.c
index 05978d82e2c7b73235c6c2244ff23846c434eba5..33cbf0499b9f0e2ae52a5b1272bfd9f0a5a027a6 100644 (file)
--- a/bacula/src/stored/spool.c
+++ b/bacula/src/stored/spool.c
@@ -248,9 +248,6 @@ static bool despool_data(DCR *dcr, bool commit)
    dcr->dblock(BST_DESPOOLING);
    dcr->despool_wait = false;
    dcr->despooling = true;
-   if (!dcr->NewVol) {
-      set_start_vol_position(dcr);
-   }
 
    /*
     * This is really quite kludgy and should be fixed some time.
@@ -281,6 +278,8 @@ static bool despool_data(DCR *dcr, bool commit)
    /* Add run time, to get current wait time */
    time_t despool_start = time(NULL) - jcr->run_time;
 
+   set_new_file_parameters(dcr);
+
    for ( ; ok; ) {
       if (job_canceled(jcr)) {
          ok = false;
@@ -301,6 +300,13 @@ static bool despool_data(DCR *dcr, bool commit)
       Dmsg3(800, "Write block ok=%d FI=%d LI=%d\n", ok, block->FirstIndex, block->LastIndex);
    }
 
+   if (!dir_create_jobmedia_record(dcr)) {
+      Jmsg(jcr, M_FATAL, 0, _("Could not create JobMedia record for Volume=\"%s\" Job=%s\n"),
+         dcr->VolCatInfo.VolCatName, jcr->Job);
+   }
+   /* Set new file/block parameters for current dcr */
+   set_new_file_parameters(dcr);
+
    /* Subtracting run_time give us elapsed time - wait_time since we started despooling */
    time_t despool_elapsed = time(NULL) - despool_start - jcr->run_time;
 

diff --git a/bacula/src/version.h b/bacula/src/version.h
index 895bbb094e8f3bf65247a2b78dad51a2dca0e0af..30f1147f9269b03572deb1c62e0ecfdcb9ec5de6 100644 (file)
--- a/bacula/src/version.h
+++ b/bacula/src/version.h
@@ -4,8 +4,8 @@
 
 #undef  VERSION
 #define VERSION "2.3.14"
-#define BDATE   "26 March 2008"
-#define LSMDATE "26Mar08"
+#define BDATE   "31 March 2008"
+#define LSMDATE "31Mar08"
 
 #define PROG_COPYRIGHT "Copyright (C) %d-2008 Free Software Foundation Europe e.V.\n"
 #define BYEAR "2008"       /* year for copyright messages in progs */

diff --git a/bacula/technotes-2.3 b/bacula/technotes-2.3
index 34f6951f76d027590b16e37e91a576761bd9a6c2..9a5fd46f5bfec4a448bd53a93eb62a42af30a12c 100644 (file)
--- a/bacula/technotes-2.3
+++ b/bacula/technotes-2.3
@@ -24,6 +24,14 @@ Add long term statistics job table
 
 
 General:
+31Mar08
+kes  Remember what volume is in VTLs (previously turned off for
+     testing).
+kes  Analysis of JobMedia records by Tom Ivar Helbekkmo <tih@hamartun.priv.no>
+     pointed out some problems in creating JobMedia records during
+     spooling with concurrent jobs. Fix the problem by updating the
+     start file position before despooling and generate a jobmedia
+     record at the end of despooling.
 28Mar08
 kes  Fix handling of JobMedia records during spooling and when a
      Volume is changed without actually writing data to the Volume.