Conference cookie::archive_backup

404.0. "unix agent dies, abs$us loops forever" by IJSAPL::KNOL (isn't every bug a bit wrong) Thu Mar 20 1997 05:28

    abs v2.0 unix save fails, abs$ubs loops forever
    customer killed the process after 10 hours
    
    The customer starts a a save of a unix client and with the 3rd thread
    he receives the following errors:
    
Executing ABS LOGIN.COM
Completed execution of ABS LOGIN.COM
"@abs_system:coordinator.com 714908FB-9BA1-11D0-A7C1-08002B398C07" Executing, output follows :
------------------------------------------------------
---------------------------------------------------------------
Starting New Request at 19-MAR-1997 23:00:35.78
   Name:   JUPITER_FULL_SAVE
   UID:    714908FB-9BA1-11D0-A7C1-08002B398C07   

COORDINATOR:  Attempting to allocate volume set AAK350...
COORDINATOR:  Retiring volume set AAK350 (exceeded consolidation criteria)
COORDINATOR:  Created new volume set: AAK321
COORDINATOR:  Mounting volume set member: AAK321 RVN 1
COORDINATOR:     (Selected drive $1$MUA11:)
COORDINATOR:  Initializing scratch volume AAK321
%MOUNT-I-MOUNTED, AAK321 mounted on _$1$MUA11: (HSJ010)
COORDINATOR:  Skipping $1$MUA11: to End of Tape...
THREAD #1:   

Operation #1 starting at 19-MAR-1997 23:03:19.82

   Data Movement Type:    FULL_SAVE
   Incremental Level:     Full Operation

   Object Set:
      Object Type:        UNIX FILES GTAR
      Include List:       /usr/*
      Exclude List:       

   Archive Information:
      Storage Class Name: UNIX_BACKUPS
      Saveset Location:   AAK321
      Saveset Name:       19MAR19972300361.

   Execution Environment:
      Name:               UNIX_BACKUPS_ENV
      Number of retries:  0
      Retry Interval:     0 minute(s)


THREAD #1:  $  
THREAD #1:  SET NOON 
THREAD #1:  $ DEFINE SYS$COMMAND sys$input: 
THREAD #1:  $ ubs := $ABS$SYSTEM:ABS$UBS.EXE  
THREAD #1:  $ ubs n?JUPITER u?"root" l?_MBA7321: - 
THREAD #1:  _$ "d?$1$MUA11:19MAR19972300361." - 
THREAD #1:  _$ af?SL 
THREAD #1:  S  s - 
THREAD #1:  _$ "c?ABSgtar -cvPGpb20 --totals --same-owner   -f - -N @1970-1-1@ /usr/*"   
THREAD #1:  %ABS-I-STARTED, ubs agent is started 
THREAD #1:       
THREAD #1:      
THREAD #1:    
THREAD #1:  Total bytes written: 442112000 
THREAD #1:  %%Status 0 
THREAD #1:  $  
THREAD #1:  Normal successful completion 
COORDINATOR:  Skipping $1$MUA11: to End of Tape...
THREAD #2:   

Operation #2 starting at 19-MAR-1997 23:14:30.99

   Data Movement Type:    FULL_SAVE
   Incremental Level:     Full Operation

   Object Set:
      Object Type:        UNIX FILES GTAR
      Include List:       /disk1/*
      Exclude List:       

   Archive Information:
      Storage Class Name: UNIX_BACKUPS
      Saveset Location:   AAK321
      Saveset Name:       19MAR19972314281.

   Execution Environment:
      Name:               UNIX_BACKUPS_ENV
      Number of retries:  0
      Retry Interval:     0 minute(s)


THREAD #2:  $  
THREAD #2:  SET NOON 
THREAD #2:  $ DEFINE SYS$COMMAND sys$input: 
THREAD #2:  $ ubs := $ABS$SYSTEM:ABS$UBS.EXE  
THREAD #2:  $ ubs n?JUPITER u?"root" l?_MBA7371: - 
THREAD #2:  _$ "d?$1$MUA11:19MAR19972314281." - 
THREAD #2:  _$ af?SL 
THREAD #2:  S  s - 
THREAD #2:  _$ "c?ABSgtar -cvPGpb20 --totals --same-owner   -f - -N @1970-1-1@ /disk1/*"   
THREAD #2:  %ABS-I-STARTED, ubs agent is started 
THREAD #2:      
THREAD #2:     
THREAD #2:    
THREAD #2:  Total bytes written: 323747840 
THREAD #2:  %%Status 0 
THREAD #2:  $  
THREAD #2:  Normal successful completion 
COORDINATOR:  Skipping $1$MUA11: to End of Tape...
THREAD #3:   

Operation #3 starting at 19-MAR-1997 23:22:17.11

   Data Movement Type:    FULL_SAVE
   Incremental Level:     Full Operation

   Object Set:
      Object Type:        UNIX FILES GTAR
      Include List:       /disk2/*
      Exclude List:       

   Archive Information:
      Storage Class Name: UNIX_BACKUPS
      Saveset Location:   AAK321
      Saveset Name:       19MAR19972322144.

   Execution Environment:
      Name:               UNIX_BACKUPS_ENV
      Number of retries:  0
      Retry Interval:     0 minute(s)


THREAD #3:  $  
THREAD #3:  SET NOON 
THREAD #3:  $ DEFINE SYS$COMMAND sys$input: 
THREAD #3:  $ ubs := $ABS$SYSTEM:ABS$UBS.EXE  
THREAD #3:  $ ubs n?JUPITER u?"root" l?_MBA7392: - 
THREAD #3:  _$ "d?$1$MUA11:19MAR19972322144." - 
THREAD #3:  _$ af?SL 
THREAD #3:  S  s - 
THREAD #3:  _$ "c?ABSgtar -cvPGpb20 --totals --same-owner   -f - -N @1970-1-1@ /disk2/*"   
THREAD #3:  %ABS-I-STARTED, ubs agent is started 
THREAD #3:      
THREAD #3:     
THREAD #3:    
THREAD #3:  %NONAME-W-NOMSG, Message number 00000000 
THREAD #3:  Agent context (subprocess) exited 
COORDINATOR:  Retiring volume set AAK321 (due to fatal error during save)
COORDINATOR:  Dismounting volume set member: AAK321 RVN 1
THREAD #3:  $  
THREAD #3:  %NONAME-W-NOMSG, Message number 00000000 
THREAD #3:  Facility ABS:  ABS_CREPRCERR, Error creating subprocess with CREPRC 
THREAD #3:     Line = 618, File = RESD$:[SRC]AGENT_SUBPROCESS.C;1 
THREAD #3:  Error creating subprocess with CREPRC 
COORDINATOR:  Invalid volume set name in Storage Class
COORDINATOR:  Facility ABS:  ABS_SLS_INVLD_VOLUME_SET, Invalid volume set name in Storage Class
COORDINATOR:     Error creating subprocess with CREPRC
COORDINATOR:  Facility ABS:  ABS_SLS_INVLD_VOLUME_SET, Invalid volume set name in Storage Class
COORDINATOR:     Line = 1209, File = RESD$:[SRC]COORD_THREAD_MANAGEMENT.C;1
COORDINATOR:  Facility ABS:  ABS_SLS_INVLD_VOLUME_SET, Invalid volume set name in Storage Class
COORDINATOR:     Line = 1903, File = RESD$:[SRC]COORD_THREAD_MANAGEMENT.C;1
COORDINATOR:  Facility ABS:  ABS_SLS_INVLD_VOLUME_SET, Invali7<d volume set name in Storage Class
COORDINATOR:     Line = 1244, File = RESD$:[SRC]COORD_ARCHIVE_MANAGEMENT.C;1
COORDINATOR:  Facility ABS:  ABS_SLS_INVLD_VOLUME_SET, Invalid volume set name in Storage Class
COORDINATOR:     Line = 1026, File = RESD$:[SRC]ARCHIVE_FILE_SYSTEM.C;1
COORDINATOR:  Facility ABS:  ABS_SLS_INVLD_VOLUME_SET, Invalid volume set name in Storage Class
COORDINATOR:     Line = 6600, File = RESD$:[SRC]SLS_SERVICES.C;1
COORDINATOR:  Facility ABS:  ABS_SLS_INVLD_VOLUME_SET, Invalid volume set name in Storage Class
COORDINATOR:     Line = 9963, File = RESD$:[SRC]SLS_SERVICES.C;1
COORDINATOR:  Final status is Invalid volume set name in Storage Class
%SYSTEM-F-OPINCOMPL, operation is incomplete
  ABS          job terminated at 20-MAR-1997 09:55:37.46

  Accounting information:
  Buffered I/O count:            5216         Peak working set size:  17504
  Direct I/O count:              2027         Peak page file size:    78960
  Page faults:                   1908         Mounted volumes:            1
  Charged CPU time:           0 00:00:16.86   Elapsed time:     0 10:55:13.39
    
    jan

1) When you say killed the process, which process was it.  Did he use
   the scheduler and issue an ABORT?  Did he STOP/ID the ABS$UBS process?

2) Has this happened more that once?  Does he have other saves with multiple
   mulitple include specs that work?

Thanks

    re.1
    
    Yes, he did a stop /id don't know if he checked the scheduler status
    of the job. I'll ask him what the scheduler output of the job is.
    
    
    The customer is trying out the abs 2.0 and came across some problems
    For as far as I know he's tried several saves already without a
    problem.
    
    
    I've set up a test with our own cluster and some unix machines but
    up to now I can't figure out which side is failing
    because a pause of the process at the unix site is seen by the abs$ubs
    and aborts , a network failure is seen and aborts. 
    The difference is that I've installed ABS V2.1 and SLS 2.9 while the
    customer is using ABS V2.0 and SLS 2.8.
    Monday I'll install ABS V2.0 and conduct some tests.
    I'll keep you posted.
    
    jan

We have uncovered a problem with ABS$UBS.EXE where UBS will hang if the
client crashes or the client "rsh" process dies without closing the
TCP/IP sockets properly.  This problem has been fixed and will be
available an ECO release of ABS V2.1.

This might be the problem that the customer has observed in this instance.

Lyle

    
    Jan,
    
    We just had this problem reported as an IPMT and sent the solution.
    The ABS$UBS.EXE image is to be placed in ABS$SYSTEM directory.
    (Make sure no unix or NT backup is active while replacing the image.)
    There is no need to restart ABS after that.  Just re-execute the unix
    backup if you want to test the image.
    
    Masami
    	

    Thanks,
    	for the quick solution, I've copied the files and will send then
    	to the customer asap.
    jan