RE: log shipping in 12c - log_archive_dest_state_2

Abhinav,

I am experiencing a similar problem in 12.1.0.2. at one of my clients. The first thought about DG problems is ALWAYS "it's the network". Usually because most of the time it's the network.

Several times a day, the physical standby (max performance mode) is lagging by up to 5 minutes and then catching up itself automatically. I am in the middle of (slowly due to other priorities) troubleshooting the problem. This clients delay seems to start with the APPLY and not the TRANSPORT, so it doesn't seem to be network. There are no obvious resource constraints.

If you have an *active* dataguard license so the standby can be open read-only during apply, you can get a lot of information by implementing Standby Statspack ( MOS: 454848.1 ) You basically create statspack accounts (_at_spcreate then _at_sbcreate) on the Primary. Add an instance (@sbaddins) and you get a database link to the standby to read the dictionary views and store them on the primary. You will need to schedule the snap's and purging yourself manually.

If you don't have OSWatcher setup, that's another good thing to have running in the background to check on resources. I'd recommend using TFA to run it (MOS: 1513912.2 ).

regards

Neil

From: iamabhinav.akg_at_gmail.com
Date: Tue, 12 Jul 2016 09:01:10 +0530
Subject: Re: log shipping in 12c - log_archive_dest_state_2 To: andrew.kerber_at_gmail.com
CC: oracle-l_at_freelists.org

We monitor network, standby, OS all the time and don't find anything mapping to this issue, so far. The pre-12c database, which exist on same set of network, OS & storage, never have this issue.

Hence, something that 12c is unable to handle or a bug.  

On Mon, Jul 11, 2016 at 7:10 PM, Andrew Kerber <andrew.kerber_at_gmail.com> wrote: It really sounds like a network problem, a network disconnect perhaps? Do you have anything monitoring the network? You could write something to monitor the standby and note when it goes out, and then identify if anything else is going on at that time.

On Mon, Jul 11, 2016 at 1:07 AM, Abhinav Gupta <iamabhinav.akg_at_gmail.com> wrote: Hi All,

I have couple of primary database (all 2 node RAC on same CRS) along with physical standby counterparts(all single instance on same server). The RDBMS versions are 12.1.0.2 and are on RHEL5.

Quite frequently one or more standby goes out of sync due to "WAIT_FOR_LOG" on mrp. This is resolved by altering log_archive_dest_state_2 on primary to DEFER and then back to ENABLE. This immediately starts the stopped log shipping and standby is back sync soon.

Mostly, one of the database is erratic, however, at times, two or more throws the same issue at same point of time. Hence, it could be something outside database as well. Never seen this issue in any of our pre-12c database, so far with similar set of configurations.

Anyone?

Regards,
Abhinav

-- 
Andrew W. Kerber

'If at first you dont succeed, dont take up skydiving.'


 		 	   		  
--
http://www.freelists.org/webpage/oracle-l