Re: slow rman compression on exadata x8-m to zfs
Date: Wed, 8 Dec 2021 13:42:42 +0000
Message-ID: <DM6PR15MB2938903F53283F74DF2E4D74A66F9_at_DM6PR15MB2938.namprd15.prod.outlook.com>
Thanks for the suggestions.
I'll do the strace and perf and also look at v$dnfs_channels a little closer.
Yeah, they're good with SR on Exadata, more so with hardware than software.
It does stink of a bug, thanks for your help
From: Mladen Gogala <gogala.mladen_at_gmail.com> Sent: December 7, 2021 9:42 PM
To: Michael McMullen <ganstadba_at_hotmail.com>; oracle-l_at_freelists.org <oracle-l_at_freelists.org> Subject: Re: slow rman compression on exadata x8-m to zfs
On 12/7/21 12:53, Michael McMullen wrote: Backups are run from a separate vm, but it doesn't matter. rman target=/ gives the same result. I've set compression on the zfs. Just trying to find a root cause.
Hi Mike,
You should also get Oracle involved and open a SR. Oracle provides very comprehensive support for their engineered systems. You should also do strace on the rman processes and see where do they spend time. You can also use perf record -p <pid> to record where the processes are spending time. It looks like the most of your rman channel processes are sleeping. 500 MB/sec is what only a single channel should be able to do. You should be able to find out by using ps -ef|grep rman Please run rman with debug=all log=/tmp/rman_debug.log parameters.
To tell the truth, this looks like a bug in rman, changing behavior when compression is used.
Regards
--
Mladen Gogala
Database Consultant
Tel: (347) 321-1217
https://dbwhisperer.wordpress.com
--
http://www.freelists.org/webpage/oracle-l
Received on Wed Dec 08 2021 - 14:42:42 CET