Oracle 12.2.0.1 seems to be the “tracing” release without further patch or parameter treatments. After posting about MMON unconditional traces in Oracle 12.2.0.1 a day ago, I received a couple of comments either on the blog, via LinkedIn or Twitter regarding other trace facilities in Oracle Database 12.2.0.1. One of them is the case that RMAN backup generates traces in Oracle 12.2.0.1.
And please, very important upfront:
Oracle Database 12.2.0.1 is a very stable and reliable release based on all the customer feedback Roy and I received so far. For instance, when we check for optimizer issues causing trouble in 12.1.0.2, most of them are fixed already in base 12.2.0.1.
Hence I disagree on some twitter comments saying “Don’t upgrade!“. This doesn’t make sense at all.
RMAN backup generates traces in Oracle 12.2.0.1
Thanks to Noons, Paul and the other people who commented and send me bug numbers and Doc-IDs.
- MOS Note: 2311975.1
RMAN BACKUP Command Always Generates Trace File Having “kcbtse_print_kod” Message
RMAN “backup” command always generates trace files containing:
kcbtse_print_kod: printing kcbtekod structure 0x7fd4ede09d1c--kcbtekod_ena = 2, kcbtekod_flag = 6 kcbtekod_mkloc = 0 --kcbtekod_key = 687c7f52909d6ff6013e45e1f4a6de6800000000000000000000000000000000 --kcbtekod_mkeyid = e097f7057c354fd8bf23f8b65d3cb9dd
The above MOS note in fact is not very helpful as it only points to the bug:
and recommends to either ignore the message or remove the trace files.
Workarounds
As the note states, there are a couple of workarounds. But one (the event) is missing (thanks Noons!):
- Remove the traces
- Apply a patch if there’s one available for your platform/update/revision release
- Set in your SPFILE:
ALTER SYSTEM SET EVENT='trace[krb.*] disk disable, memory disable' COMMENT='RMAN Bugs 26844870 and 22700845 - set May 24, 2018' SCOPE=SPFILE;
The event will suppress the traces. BUT … it will ONLY work if you are at least on the April 2018 Update as otherwise this bug hits you:
and the event won’t have an effect. The fix for Bug 22700845 is included in the July 2018 Update, the April 2018 Windows BP and in Oracle 18.1.0. In releases older than that – in my understanding – the event won’t have an effect and you’ll have to rely on cleaning out the traces.
You’ll have to restart your database to have the above event take effect.
And finally, if I forgot something, please let me know.
Further Information
You may also have a look at:
–Mike
YAY! Thanks heaps for this, Mike!
I still have not had a soution to the SR I raised on this.
But as you state, it is fixed in the July 2018 RU, going to download and install that one straight away!
I’m dying to use the new online maintenance features of 12.2 but this issue with RMAN put the brakes on that.
No more! 12.2, here I come!
This has now been fixed by Bug 28390273 – RMAN: KRB traces in trace directory on running every command (Doc ID 28390273.8)
Hi,
This issue is happening on 12.1.0.2 databases as well. Will setting the event work for 12.1 databases?
Thanks,
Roya,
I think so – but PLEASE open an SR. I have seen the issue shown to me at a customer’s system yesterday at the AOUG conference in Vienna – on 18.6.0. So the issues seems to be back again. The event should work in 12.1.0.2 as well I guess.
Cheers,
Mike
Take a look at bug 29061016: Huge tracefile generated by RMAN incremental Backups
I’ve witnessed personally on a customer running 12.2.0.1.191015
The workaround on the spfile worked for me.