Re: CPU waiting for I/O in Exadata Database Node
From: Ls Cheng <exriscer_at_gmail.com>
Date: Tue, 22 May 2018 23:36:51 +0200
Message-ID: <CAJ2-Qb9F2FFaeODtuXn9_ODtG620pa75rvYL6c33hdC1QEX91Q_at_mail.gmail.com>
Date: Tue, 22 May 2018 23:36:51 +0200
Message-ID: <CAJ2-Qb9F2FFaeODtuXn9_ODtG620pa75rvYL6c33hdC1QEX91Q_at_mail.gmail.com>
Hi Steve
Thanks for the tip, there are indeed several NFS mounts.
Thanks
On Tue, May 22, 2018 at 10:16 PM, Steve Harville <steve.harville_at_gmail.com> wrote:
> I have seen this on Exadata when reading/writing to NFS mounts.
>
> On Tue, May 22, 2018 at 1:29 PM Ls Cheng <exriscer_at_gmail.com> wrote:
>
>> Hi
>>
>> I am observing quite high CPU waiting for I/O in Exadata Database Nodes
>> (vmstat output). but from iostat there no waits at all (which is obvious
>> since heavy I/O is done through infiniband).
>>
>> Anyone know what can be the CPU waiting for?
>>
>> Thanks
>>
>>
>>
-- http://www.freelists.org/webpage/oracle-lReceived on Tue May 22 2018 - 23:36:51 CEST