[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference cookie::smfs

Title:Sequential Media Filesystem for OpenVMS VAX
Moderator:COOKIE::KYLER
Created:Mon Aug 30 1993
Last Modified:Thu May 01 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:66
Total number of notes:198

65.0. "SMFS not unloading tapes." by CSC32::BSS::SAUL () Mon Mar 03 1997 13:58

Customer mounts and copies to 3 files systems.  First one dismounts and 
unloads ok but the 2nd and 3rd only dismount.  They never unload.  Think
of a reason why?  Also, customer wants to know if there is a way to limit
the number of drives that SMFS uses.  She has a robot with four drives but
only wants to SMFS to access two of them.  Thanks,

Ted
T.RTitleUserPersonal
Name		DateLines
65.1CSC32::BSS::SAULMon Mar 03 1997 14:586
Also, the two drives are in the drives still if there is any 
information you would like gathered from the customer.  We have
until 16:30 Mountain time before the brute force method will have
to be used to clear the drives.

Ted
65.2COOKIE::KYLERTue Mar 04 1997 12:4513
    SMF intentionally leaves the tape in the drive as an optimization,
    presuming that it is likely to need it again.  Why is this a problem? 
    Other cooperating (i.e. using MDMS) applications are free to use the
    drive.  A load request will cause the tape in the drive to be unloaded.
    
    >Also, customer wants to know if there is a way to limit the number of 
    >drives that SMFS uses.  She has a robot with four drives but
    >only wants to SMFS to access two of them.
    
    The only way to do this effectively is by giving the drives different
    media types.
    
    - Dan.
65.3CSC32::BSS::SAULTue Mar 04 1997 17:2414
    More information:
    
    I found out that the drives were not only loaded but still allocated to
    SMF.  Next note will be the output of the show device on the two
    drives.  This isn't normal is it?
    
    The docs say on page 6-3 that you can assign and dedicate if desired
    any number of devices to SMF.  Is this referring to the solution you
    gave?
    
    Ted
65.4CSC32::BSS::SAULTue Mar 04 1997 17:2761
VAX_UHAD8$  robot show robot  $1$dua4001: drive
DRIVE:          0               TLB015
DRIVE:          1               TLB013
DRIVE:          2               Empty
DRIVE:          3               Empty


Magtape $200$MUA120: (HSJ305), device type TZ88, is online, allocated,
    controller supports tape data caching (write-back cache enabled), mounted,
    volume is marked for dismount, file-oriented device, served to cluster via
    TMSCP Server, error logging is enabled.

    Error count                    0    Operations completed
29
    Owner process          "MXD1ACP"    Owner UIC
[VMS,SYSTEM]
    Owner process ID        2020231B    Dev Prot
S:RWPL,O:RWPL,G:R,W:RWPL
    Reference count                2    Default buffer size
10000

    Volume label            "TLB015"    Relative volume no.
0
    Record size                    0    Transaction count
2
    Mount status             Process    Mount count
1
    ACP process name       "MXD1ACP"
    Density                     6666    Format
Normal-11
    Host name               "HSJ305"    Host type, avail              HSJ3,
yes
    Alternate host name     "HSJ306"    Alt. type, avail              HSJ3,
yes
    Allocation class             200

  Volume status:  no-unload on dismount, odd parity.

Magtape $200$MUA130: (HSJ306), device type TZ88, is online, allocated,
    controller supports tape data caching (write-back cache enabled), file-
    oriented device, served to cluster via TMSCP Server, error logging is
    enabled.

    Error count                    0    Operations completed
94
    Owner process          "MXU1ACP"    Owner UIC
[VMS,SYSTEM]
    Owner process ID        20202317    Dev Prot
S:RWPL,O:RWPL,G:R,W:RWPL
    Reference count                1    Default buffer size
10000
    Density                     6666    Format
Normal-11
    Host name               "HSJ306"    Host type, avail              HSJ3,
yes
    Alternate host name     "HSJ305"    Alt. type, avail              HSJ3,
yes
    Allocation class             200

  Volume status:  no-unload on dismount, beginning-of-tape, odd parity.
65.5COOKIE::KYLERThu Mar 06 1997 15:1416
    >I found out that the drives were not only loaded but still allocated to
    >SMF.  Next note will be the output of the show device on the two
    >drives.  This isn't normal is it?
    
    It is not normal unless there is still a file open.  Some information
    which may help would be the output of SDA SHOW DEVICE on the tape
    drives, and SDA SHOW PROCESS/CHANNELS on the process which created the
    files.
    
    >The docs say on page 6-3 that you can assign and dedicate if desired
    >any number of devices to SMF.  Is this referring to the solution you
    >gave?
    
    I believe it must be, because that's the only way I know of to do it.
    
    - Dan.
65.6From the customer..ignore the error count. They came from a TL812 reset.CSC32::BSS::SAULMon Mar 10 1997 10:23334


VAX_UHAD8$  sh dev      mua130

Device                  Device           Error    Volume         Free  Trans
Mnt
 Name                   Status           Count     Label        Blocks Count
Cnt
$200$MUA130:  (HSJ305)  Mounted dismount     3     TLB014            0     2
1
                        cached alloc



>SDA>SHOW DEVICE on_drive_thats_hung

I/O data structures
-------------------
(HSJ306$MUA130) HSJ305$MUA130           TZ88              UCB address:
8896E9C0

Device status:   00000810 online,valid
Characteristics: 0CEC4038 dir,sdi,sqd,fod,avl,mnt,dmt,elg,all,idv,odv
                 000026B1 clu,2p,mscp,srv,nnm,wbc,loc

Owner UIC [000001,000004]   Operation count         22   ORB address
88539C80
      PID        0001003F   Error count              3   DDB address
88251240
Alloc. lock ID   08008F63   Reference count          2   DDT address
88FAA794
Alloc. class          200   BOFF                  0000   VCB address
889AB740
Class/Type          02/3E   Byte count            0000   CRB address
882512C0
Def. buf. size      32768   SVAPTE            00000000   PDT address
880D6B90
DEVDEPEND        000011C0   DEVSTS                4030   CDDB address
88D95700
DEVDEPND2        001102FC   RWAITCNT              0000   2P_CDDB addr.
88D94800
FLCK index             34   Object count            41   2P_DDB address
88249AC0
DLCK address     00000000                                I/O wait queue
empty
Device   DEVSTS   status:   4030 bit4,bit5,mscp_ignsrv

I/O data structures
-------------------

                --- Primary Class Driver Data Block (CDDB) 88D95700 ---

Status:              0040 alcls_set
Status2:             0000
Controller Flags:    00D4 cf_mlths,cf_this,cf_misc,cf_attn

Allocation class     200    CDRP Queue         empty    DDB address
88251240
System ID       100A1720    Restart Queue      empty    CRB address
882512C0
                    4200    DAP Count              2    CDDB link
88D95C00
Contrl. ID      51900678    Contr. timeout       255    PDT address
880D6B90
                01270009    Reinit Count           0    Original UCB
00000000
Response ID     00000000    Wait UCB Count         0    UCB chain
88973600
MSCP Cmd status 00000000

I/O data structures
-------------------

                --- Secondary Class Driver Data Block (CDDB) 88D94800 ---

Status:              0040 alcls_set
Status2:             0000
Controller Flags:    00D4 cf_mlths,cf_this,cf_misc,cf_attn

Allocation class     200    CDRP Queue         empty    DDB address
88249AC0
System ID       10041120    Restart Queue      empty    CRB address
88249B40
                    4200    DAP Count              1    CDDB link
88D94D00
Contrl. ID      42700301    Contr. timeout       255    PDT address
880D6B90
                01270001    Reinit Count           0    Original UCB
00000000
Response ID     00000000    Wait UCB Count         0    UCB chain
8896EB00
MSCP Cmd status 00000000

        *** I/O request queue is empty ***


I/O data structures
-------------------

                --- Volume Control Block (VCB) 889AB740 ---

Volume: TLB014
Status:   00
Mode:   0000

Transactions           2    Rel. volume            0    AQB address
881BF7C0
Start record          36    Tape vol. list  881C4340    Virt. pg. queue
00279270
Tapemark count         0                                Blocked queue
empty

I/O data structures
-------------------

                    --- ACP Queue Block (AQB) 881BF7C0 ---

ACP requests are serviced by process MXS1ACP whose PID is 0001003F

Status: 00

Mount count            1    ACP type       undefined    Linkage
881BAEC0
                            ACP class            136    Request queue
00000000

        *** ACP request queue is empty ***

I/O data structures
-------------------
HSJ305$MUA130 (HSJ306$MUA130)           TZ88              UCB address:
8896E9C0

Device status:   00000810 online,valid
Characteristics: 0CEC4038 dir,sdi,sqd,fod,avl,mnt,dmt,elg,all,idv,odv
                 000026B1 clu,2p,mscp,srv,nnm,wbc,loc

Owner UIC [000001,000004]   Operation count         22   ORB address
88539C80
      PID        0001003F   Error count              3   DDB address
88251240
Alloc. lock ID   08008F63   Reference count          2   DDT address
88FAA794
Alloc. class          200   BOFF                  0000   VCB address
889AB740
Class/Type          02/3E   Byte count            0000   CRB address
882512C0
Def. buf. size      32768   SVAPTE            00000000   PDT address
880D6B90
DEVDEPEND        000011C0   DEVSTS                4030   CDDB address
88D95700
DEVDEPND2        001102FC   RWAITCNT              0000   2P_CDDB addr.
88D94800
FLCK index             34   Object count            41   2P_DDB address
88249AC0
DLCK address     00000000                                I/O wait queue
empty
Device   DEVSTS   status:   4030 bit4,bit5,mscp_ignsrv

I/O data structures
-------------------

                --- Primary Class Driver Data Block (CDDB) 88D95700 ---

Status:              0040 alcls_set
Status2:             0000
Controller Flags:    00D4 cf_mlths,cf_this,cf_misc,cf_attn

Allocation class     200    CDRP Queue         empty    DDB address
88251240
System ID       100A1720    Restart Queue      empty    CRB address
882512C0
                    4200    DAP Count              1    CDDB link
88D95C00
Contrl. ID      51900678    Contr. timeout       255    PDT address
880D6B90
                01270009    Reinit Count           0    Original UCB
00000000
Response ID     00000000    Wait UCB Count         0    UCB chain
88973600
MSCP Cmd status 00000000


I/O data structures
-------------------

                --- Secondary Class Driver Data Block (CDDB) 88D94800 ---

Status:              0040 alcls_set
Status2:             0000
Controller Flags:    00D4 cf_mlths,cf_this,cf_misc,cf_attn

Allocation class     200    CDRP Queue         empty    DDB address
88249AC0
System ID       10041120    Restart Queue      empty    CRB address
88249B40
                    4200    DAP Count              1    CDDB link
88D94D00
Contrl. ID      42700301    Contr. timeout       255    PDT address
880D6B90
                01270001    Reinit Count           0    Original UCB
00000000
Response ID     00000000    Wait UCB Count         0    UCB chain
8896EB00
MSCP Cmd status 00000000

        *** I/O request queue is empty ***


I/O data structures
-------------------

                --- Volume Control Block (VCB) 889AB740 ---

Volume: TLB014
Status:   00
Mode:   0000

Transactions           2    Rel. volume            0    AQB address
881BF7C0
Start record          36    Tape vol. list  881C4340    Virt. pg. queue
00279270
Tapemark count         0                                Blocked queue
empty


I/O data structures
-------------------

                    --- ACP Queue Block (AQB) 881BF7C0 ---

ACP requests are serviced by process MXS1ACP whose PID is 0001003F

Status: 00

Mount count            1    ACP type       undefined    Linkage
881BAEC0
                            ACP class            136    Request queue
00000000

        *** ACP request queue is empty ***



>SDA>SHOW PROCESS/CHAN on_process_thats_creating_files

Process index: 0167   Name: Hyland on VT222   Extended PID: 20401167
--------------------------------------------------------------------


                            Process active channels
                            -----------------------

Channel  Window           Status        Device/file accessed
-------  ------           ------        --------------------
  0010  00000000                        DSA0:
  0020  8821CE80                        DSA0:[VMS$COMMON.SYSEXE]SDA.EXE;1
  0030  8819DE00
DSA0:[VMS$COMMON.SYSLIB]SECURESHRP.EXE;2
 (section file)
  0040  00000000                        VTA222:
  0050  8819FD00
DSA0:[VMS$COMMON.SYSLIB]SECURESHR.EXE;2
(section file)
  0060  00000000                        VTA222:
  0070  881A1700                        DSA0:[VMS$COMMON.SYSLIB]LIBRTL.EXE;14
(s
ection file)
  0080  8819F0C0                        DSA0:[VMS$COMMON.SYSLIB]SMGSHR.EXE;2
(se
ction file)
  0090  8819BA80                        DSA0:[VMS$COMMON.SYSEXE]DCL.EXE;1
(secti
    Press RETURN for more.
SDA> 0  8819EF40                        DSA0:[VMS$COMMON.SYSLIB]SCRSHR.EXE;1
(se
ction file)
Process index: 0167   Name: Hyland on VT222   Extended PID: 20401167
--------------------------------------------------------------------

Channel  Window           Status        Device/file accessed
-------  ------           ------        --------------------
  00B0  881B47C0
DSA0:[VMS$COMMON.SYSLIB]DCLTABLES.EXE;26
8 (section file)
  00C0  8819EAC0                        DSA0:[VMS$COMMON.SYSLIB]LBRSHR.EXE;14
(s
ection file)
  00D0  00000000                        VTA222:
  00E0  8865E6C0                        DSA0:[UHCOMMON.VICKIE.LIB]SDA.INIT;1
  0100  00000000                        VTA222:
  0110  00000000                        VTA222:
  0120  881A0700
DSA0:[VMS$COMMON.SYSMSG]PRGDEVMSG.EXE;1
(section file)
  0130  881A0880
DSA0:[VMS$COMMON.SYSMSG]SHRIMGMSG.EXE;1
(section file)



VAX_UHAD8$  sh dev/full mua130

Magtape $200$MUA130: (HSJ305), device type TZ88, is online, allocated,
    controller supports tape data caching (write-back cache enabled), mounted,
    volume is marked for dismount, file-oriented device, served to cluster via
    TMSCP Server, error logging is enabled.

    Error count                    3    Operations completed
22
    Owner process          "MXS1ACP"    Owner UIC
[VMS,SYSTEM]
    Owner process ID        2040023F    Dev Prot
S:RWPL,O:RWPL,G:R,W:RWPL
    Reference count                2    Default buffer size
32768

    Volume label            "TLB014"    Relative volume no.
0
    Record size                    0    Transaction count
2
    Mount status             Process    Mount count
1
    ACP process name       "MXS1ACP"
    Density                     6666    Format
Normal-11
    Host name               "HSJ305"    Host type, avail              HSJ3,
yes
    Alternate host name     "HSJ306"    Alt. type, avail              HSJ3,
yes
    Allocation class             200

  Volume status:  no-unload on dismount, odd parity.
65.7CSC32::BSS::SAULThu Mar 13 1997 09:461
Any ideas?  This is a mission critical customer.
65.8COOKIE::KYLERMon Mar 17 1997 15:414
    Everything here looks reasonably normal, except for the fact that the
    device is not being dismounted.   Can the MX device be SMF DISMOUNTed? 
    If so, does that clear the tape device?
65.9CSC32::BSS::SAULWed Mar 19 1997 10:2018
>    Everything here looks reasonably normal, except for the fact that the
>    device is not being dismounted.   Can the MX device be SMF DISMOUNTed? 
>    If so, does that clear the tape device?
 
Yes it does, but these are manual steps that she has to take to clear the drives.

Regarding SMF not unloading the tapes, customer has these comments:
 
"SMF has never unloaded the
drives.  In a test I performed SMF confiscated all 4 tape drives, and
backups failed with the SLS message "no tape drives available to
allocate" or something similar to that.

I need to use SMF.  But it must unload the drives for security reasons
when finished with the volume; data security is more important than any
performance goals.  And it absolutely must unload the drives when they
are needed by other applications.  These parameters should be set by the
customer, like HSM does."
65.10COOKIE::KYLERWed Mar 19 1997 16:5920
>Yes it does, but these are manual steps that she has to take to clear the drives.
    
    I understand that.  I asked the question to help in my diagnosis, not
    to suggest a workaround.
    
    SMF DISMOUNT causes execution of the same code which should be executed
    about 30 seconds after the file is closed.  The fact that the dismount
    succeeds and clears the drive tells me that no file is open, and that
    the problem is probably with the timer, not with the code it executes.
    
    I recall that there was a problem with pthread_delay (cma_delay) where
    for small time values, it never returned.  This would cause the
    behavior we're seeing, but I'm skeptical of this, because the SMF delay
    involved is 15 seconds, and this usually happened with delays of one
    second or less.  There is a STARs article and a VMS patch for this
    problem.
    
    Could I get VMS, MDMS, and SMF versions and architectures?
    
    - Dan.
65.11CSC32::V_HEINICKEThu Apr 03 1997 13:2317
    Hi,
    
    Ted will be out of the office until next week.  
    
    Reviewing the call indicates that the customer is currently running the
    following:
    
    SMFS V1.2
    OpenVMS V6.2 (VAX)
    OpenVMS V6.2-1h2 (ALPHA)
    SLS/MDMS  V2.8 ECO2
    
    Customer just came back in stating SMF is still unusable.  
    
    Thanks,
    
    Victoria
65.12COOKIE::KYLERFri Apr 04 1997 16:483
    How about the CMA patch?  Do they have it, if it is applicable to VMS
    6.2?
65.13CSC32::BSS::SAULMon Apr 28 1997 12:274
The patch is VAXCMAR04_062.  I did not find an AXP version of the patch.
I have asked the customer if this has been applied or not.

Ted
65.14CSC32::BSS::SAULMon Apr 28 1997 12:446
Customer has both VAXCMAR03_062 and ALPCMAR04_062 installed.

She is pushing on this so I think I will have to do an IMPT to
expedite this.

Ted