Disk File Optimizer Error

[Lisa Collins]

I did what the document showed and this is what I see

DFO>sh DFG$SCRIPT32

File script DFG$SCRIPT32 is ENABLED, LEVEL=1, device: NEOMN2$DRA1
No defragmentation process is executing
Scheduler queue contains no pending operations
File list for DEFRAGMENT FILE:
[CSWS.APACHE.SPECIFIC.NEOMN2.NEOMIN.ORG.LOGS]*.*;*
Do not consolidate freespace
Minimum Priority: 3
Default Priority: 4
Script may operate on any available node
No log file specified
No prologue file specified
No epilogue file specified
Mail notification is OFF
OPCOM notification is ON
Special file handling:
Exclude all specifically placed files

DFO>

It does not include a log like the example showed in the document.


We rebooted the server this morning and things seemed fine for 45 minutes. Then the server rebooted itself. I don't know if anyone can make heads or tales out of my dump file but I'll post it just incase



$ ANALYZE/CRASH_DUMP SYS$SYSTEM:SYSDUMP.DMP

OpenVMS (TM) Alpha system dump analyzer
...analyzing a compressed selective memory dump...

Dump taken on 3-MAR-2006 07:26:28.09
CPUSPINWAIT, CPU spinwait timer expired

SDA> sh process

Process index: 0140 Name: PMDFF9 pmas Extended PID: 00000540
--------------------------------------------------------------------
Process status: 00040001 RES,PHDRES
status2: 00000001 QUANTUM_RESCHED

PCB address 8170AF80 JIB address 81705BC0
PHD address 85E28000 Swapfile disk address 00000000
KTB vector address 8170B26C HWPCB address FFFFFFFF.85E28080
Callback vector address 00000000 Termination mailbox 001D
Master internal PID 00010140 Subprocess count 0
Creator extended PID 00000000 Creator internal PID 00000000
Previous CPU Id 00000000 Current CPU Id 00000001
Previous ASNSEQ 000000000000034A Previous ASN 0000000000000011
Initial process priority 4 # open files remaining 193/200
Delete pending count 0 Direct I/O count/limit 200/200
UIC [00001,000004] Buffered I/O count/limit 10000/10000
Abs time of last event 00025111 BUFIO byte count/limit 98528/98528
# of threads 1 ASTs remaining 248/250
Swapped copy of LEFC0 00000000 Timer entries remaining 199/200
Swapped copy of LEFC1 00000000 Active page table count 0
Global cluster 2 pointer 00000000 Process WS page count 526

Press RETURN for more.
SDA>

Process index: 0140 Name: PMDFF9 pmas Extended PID: 00000540
--------------------------------------------------------------------
Global cluster 3 pointer 00000000 Global WS page count 168
PCB Specific Spinlock 81709AC0


Press RETURN for more.
SDA>

Process index: 0140 Name: PMDFF9 pmas Extended PID: 00000540
--------------------------------------------------------------------

Thread index: 0000
------------------
Current capabilities: System: 0000000C QUORUM,RUN
User: 00000000
Permanent capabilities: System: 0000000C QUORUM,RUN
User: 00000000
Current affinities: 00000000
Permanent affinities: 00000000
Thread status: 00040001
status2: 00000001

KTB address 8170AF80 HWPCB address FFFFFFFF.85E28080
PKTA address 7FFEFF98 Callback vector address 00000000
Internal PID 00010140 Callback error 00000000
Extended PID 00000540 Current CPU id 00000001
State CUR 01 Flags 00000000
Base priority 4 Current priority 4
Waiting EF cluster 4 Event flag wait mask 8100C100

Press RETURN for more.
SDA>

Process index: 0140 Name: PMDFF9 pmas Extended PID: 00000540
--------------------------------------------------------------------
CPU since last quantum FFF7 Mutex count 0
ASTs active NONE

[Volker Halle]

Lisa,

you would have to specify a logfile for the script to receive more detailled error messages.

To be able to help you regarding the CPUSPINWAIT system crash, please attach the CLUE file from CLUE$COLLECT:CLUE$node_ddmmyy_hhmm.LIS from today as a .TXT attachment.

Volker.

[Ian Miller.]

What version of VMS are you running?

Is there a file in SYS$ERRORLOG or SYS$MANAGER with a name starting with CLUE dated when the system crashed ? If so can you post it as an attachment.

[Lisa Collins]

This is our system

AlphaServer DS20 500 MHz running OpenVMS Alpha V7.3-1

[Zeni B. Schleter]

As of OpenVMS v7.1 there were references in patches about spinwait Bugchecks. Prior to patches, a workaround was setting the SMP_SPINWAIT to 300000 on those systems with multiple CPUs. We never set the SMP_SPINWAIT back down but we have not seen any problems with SPINWAIT crashes either.

The v7.1 Driv08 patch specifically referred to setting the SMP_SPINWAIT to 300000.

[Zeni B. Schleter]

I left out that I was talking about the SYSGEN parameter of SMP_SPINWAIT.

[Lisa Collins]

That's OK, I figured it was in sysgen. Currently that parameter is set to 100000 and it does not seem to be a dynamic parameter. Looks like another reboot is in order.

Another question that some of you might be able to give me advice on is should I have concerns about running DFO on a very active disk. The server I am dealing with is an email server. So even if I run the optimizing software at night, certain disks will be fairly active. (Ex: one disk runs anti-virus software, another disk runs the anti-spam software, etc.) In my reading, I had not read that it is necessary to run the DFO software with certain services shut down. But I thought that some of you with experience using DFO may have an opinion on this.

Thanks, Lisa

[Phillip Thayer]

Lisa,

Zeni is correct. There was a bug with V7.3-1 where Spinwait was set to low and you had to set it to 300000 or you would suffer from performance problems and system crashes. I have run into this as well.

Set it using the MIN_SMP_SPINWAIT in modparams and autogen with a reboot.

Phil

[Lisa Collins]

Also, this document

http://mvb.saic.com/disk$vmsdoc0731/000000/731final/6652/6652pro_009.html

refers to the SMP_LNGSPINWAIT system parameter.

[Phillip Thayer]

Lisa,

On the second part of your post about running DFO on a busy disk. You can do this but it will take longer to defragment the disk. Once you get the DFO working correctly, you can set it up so that it runs all the time with I/O Trhottled back so it will not interfere with the applications running on your disk. Not sure if DFO has this but I know Diskeeper does so I'm guessing that DFO does have I/O Throttling.

Phil