All these symptoms are consistent with a case of 'lost CDS', where XCF has
determined for whatever reason that the missing CDS is 'defective'. I
haven't seen this condition for years, but it has happened--and not always
traceable to a clear cause. The number one suspect is volume RESERVE. If
XCF cannot complete I/O to a CDS for even a short period, the dataset will
be declared unusable and removed from the configuration. When this
situation occurs, XCF puts out a complaining or two and rolls on. Like
having your spare tire go flat. You probably won't notice until you have to
move the bodies in the trunk to pull the spare out. If you have all syslog
since the previous IPL, you will most certainly find a flurry of IX*
messages chronicling the CDS removal.
You might not notice any of this until the next IPL. Years ago we created
an automation routine to notify us loudly and clearly that we had lost a
CDS. I forget the message(s) we key on, but I could dig into it if
requested. The last episode probably predated Health Checker. The current
(1.9) HC does not seem to specifically list 'CDS consistency', but there
are a couple of XCF checks that I would expect to holler if a named pair
was not intact.
Meanwhile, make sure that all CDSs live on their own volumes. It seems
wasteful to dedicate that much space to a handful of smallish data sets,
but the effects of a sysplex wide outage far outweigh the unused cylinders.
Also watch out for full volume backups that might grab a volume for an
extended period. Better not to back up CDS volumes at all.
.
.
JO.Skip Robinson
Southern California Edison Company
Electric Dragon Team Paddler
SHARE MVS Program Co-Manager
626-302-7535 Office
323-715-0595 Mobile
[EMAIL PROTECTED]
"McKown, John"
<[EMAIL PROTECTED]
THMARKETS.COM> To
Sent by: IBM [email protected]
Mainframe cc
Discussion List
<[EMAIL PROTECTED] Subject
.EDU> couple dataset problem at IPL?
08/04/2008 06:00
AM
Please respond to
IBM Mainframe
Discussion List
<[EMAIL PROTECTED]
.EDU>
This is the first time that this has happened and I am confused. What
would make a couple dataset "go bad"? We have a two system, basic
sysplex. This weekend, we IPL'ed both systems. Our procedure is to take
down SY1 first, and reIPL it. When SY1 is up, then shutdown SY2 and
reIPL. On the IPL of SY1, we got messages:
IXC268I THE COUPLE DATA SETS SPECIFIED IN COUPLE11 ARE IN AN INCONSISTEN
T STATE
IXC275I COUPLE DATA SETS SPECIFIED IN COUPLE11 ARE 168
PRIMARY: SYS1.LIH1.XCFCDS01
ON VOLSER LIHTS4
ALTERNATE: SYS1.LIH1.XCFCDS02
ON VOLSER LIHTS6
IXC273I XCF ATTEMPTING TO RESOLVE THE COUPLE DATA SETS
IXC275I RESOLVED COUPLE DATA SETS ARE 170
PRIMARY: SYS1.LIH1.XCFCDS01
ON VOLSER LIHTS4
ALTERNATE:
NONE USED
I did a SETXCF to make SYS1.LIH1.XCFCDS02 the alternate.
SETXCF COUPLE,ACOUPLE=SYS1.LIH1.XCFCDS02
IXC309I SETXCF COUPLE,ACOUPLE REQUEST FOR SYSPLEX WAS ACCEPTED
IXC260I ALTERNATE COUPLE DATA SET REQUEST FROM SYSTEM 583
LIH1 FOR SYSPLEX IS NOW BEING PROCESSED.
DATA SET: SYS1.LIH1.XCFCDS02
IEF196I IEF237I 2103 ALLOCATED TO SYS00099
IXC251I NEW ALTERNATE DATA SET 585
SYS1.LIH1.XCFCDS02
FOR SYSPLEX HAS BEEN MADE AVAILABLE
On SY2 at IPL, we got:
*IXC267E PROCESSING WITHOUT AN ALTERNATE 801
COUPLE DATA SET FOR SYSPLEX.
ISSUE SETXCF COMMAND TO ACTIVATE A NEW ALTERNATE.
*IXC267E PROCESSING WITHOUT AN ALTERNATE 802
COUPLE DATA SET FOR WLM.
ISSUE SETXCF COMMAND TO ACTIVATE A NEW ALTERNATE.
IEF196I IEF237I 2802 ALLOCATED TO SYS00003
*IXC267E PROCESSING WITHOUT AN ALTERNATE 803
COUPLE DATA SET FOR LOGR.
ISSUE SETXCF COMMAND TO ACTIVATE A NEW ALTERNATE.
I issued the SETXCF commands to activate the alternates, and got the
message:
SETXCF COUPLE,TYPE=LOGR,ACOUPLE=SYS1.LIH1.LOGRCD01
IXC309I SETXCF COUPLE,ACOUPLE REQUEST FOR LOGR WAS ACCEPTED
IXC260I ALTERNATE COUPLE DATA SET REQUEST FROM SYSTEM 750
DEV1 FOR LOGR IS NOW BEING PROCESSED.
DATA SET: SYS1.LIH1.LOGRCD01
IEF196I IEF237I 2103 ALLOCATED TO SYS00100
IXC248E COUPLE DATA SET 752
SYS1.LIH1.LOGRCD01 ON VOLSER LIHTS6
FOR LOGR MAY BE IN USE BY ANOTHER SYSPLEX.
28 IXC247D REPLY U TO ACCEPT USE OR D TO DENY USE OF THE COUPLE DATA SET
FOR LOGR.
IEE600I REPLY TO 28 IS;U
IXC251I NEW ALTERNATE DATA SET 757
SYS1.LIH1.LOGRCD01
FOR LOGR HAS BEEN MADE AVAILABLE
Any ideas?
--
John McKown
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
