Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

rollforward ignoring overflow log path?

420 views
Skip to first unread message

Lennart Jonsson

unread,
Aug 22, 2014, 10:32:40 PM8/22/14
to

I restored a database and retrieved logfiles which where placed in a
directory, say /db2/db2extractlogs/${db}. The logfiles are :

S0144485.LOG - S0144520.LOG

Now, when I do the rollforward:

db2 "rollforward db ${db} to 2014-07-08-18.10.03 using local time
overflow log path ( /db2/db2extractlogs/${db} )"

SQL4970N Roll-forward recovery on database "NYA" cannot reach the
specified stop point (end-of-log or point-in-time) on database
partition(s) "0". Roll-forward recovery processing has halted on log
file "S0144486.LOG".


I tried all kind of stuff until I finally gave up and copied the
logfiles to the directory pointed to by "Path to log files" in db cfg

Now when I rerun the exact same roll-forward as before it succeeds:

Rollforward Status

Input database alias = nya
Number of nodes have returned status = 1

Node number = 0
Rollforward status = DB working
Next log file to be read = S0144514.LOG
Log files processed = S0144485.LOG - S0144513.LOG
Last committed transaction = 2014-07-08-18.09.41.000000 Local


What could be the reason that db2 does not accept log-files from the
directory pointed to by overflow logpath? Could it be a problem if the
overflow directory resides on another partition (I haven't verified
whether that is the case)?




db2level
DB21085I This instance or install (instance name, where applicable:
"db2inst1") uses "64" bits and DB2 code release "SQL09078" with level
identifier "08090107".
Informational tokens are "DB2 v9.7.0.8", "s130316", "IP23433", and Fix Pack
"8".
Product is installed at "/opt/ibm/db2/V9.7".

Lennart Jonsson

unread,
Aug 22, 2014, 10:36:43 PM8/22/14
to
On 08/23/2014 04:32 AM, Lennart Jonsson wrote:
[...]
>Could it be a problem if the
> overflow directory resides on another partition

meaning disk partition, another fs

Rahul Chaukse

unread,
Nov 29, 2014, 2:55:20 AM11/29/14
to
before rolling forward do not forget to zip or delete the older active logs i.e. the logs mentioned in "Path to log files"

Luiz da Silva

unread,
Nov 29, 2014, 6:02:01 PM11/29/14
to
I think the default is UTC Time... Have you tried that instead of Local Time?

Peter H. Coffin

unread,
Nov 29, 2014, 9:25:06 PM11/29/14
to
On Sat, 29 Nov 2014 15:01:58 -0800 (PST), Luiz da Silva wrote:
> On Friday, August 22, 2014 11:32:40 PM UTC-3, Lennart Jonsson wrote:
>>
>> db2 "rollforward db ${db} to 2014-07-08-18.10.03 using local time
>> overflow log path ( /db2/db2extractlogs/${db} )"
>>
[...]
>
> I think the default is UTC Time... Have you tried that instead of Local Time?

And I've *usually* seen the timestamps for the rolls and backups using
no delimiters: 20140708181003

--
A *huge* proportion of people cannot make *correct and accurate*
generalisations of principles. They have to learn everything as if
it's an unrelated piece of crap, BECAUSE THEY ARE STUPID! PEOPLE ARE
STUPID! -- Thorfinn in the Monastery

jba...@calculo-sa.es

unread,
Dec 1, 2014, 4:08:51 AM12/1/14
to
How was your RESTORE command?

We use this format:

time db2 restore db mydb from '/backups' taken at $date on '/data/tbsppre' into newdb LOGTARGET '/logs/logtarget' newlogpath '/logs/pre' without prompting

and this is our rollforward command:

time db2 rollforward db newdb to end of logs and stop overflow log path '(/logs/logtarget)'
0 new messages