RE: parallel recovery slaves waiting on undo reads

From: Noveljic Nenad <nenad.noveljic_at_vontobel.com>
Date: Wed, 26 Feb 2020 23:49:38 +0000
Message-ID: <39394_1582760994_5E570421_39394_6021_1_e3f8bfab30d743f3b4376cd64610242f_at_vontobel.com>



>is this Active Data Guard?

No.

> It’s probably worth seeing if you can catch individual events in v$session_wait_history (shows the last 10 events per session). Importantly, see if the longer durations match up to more requests in the same

Yes, that’s exactly the case! Well guessed, Andy! Thanks for mentioning v$session_wait_history – I’ve never used it until now. Oracle database always pleasantly surprises with the sophistication of its instrumentation.

SQL> select event,p2,p2text from v$session_Wait_history where sid=545 ;

EVENT                                                                    P2 P2TEXT
---------------------------------------------------------------- ---------- ----------------------------------------------------------------
db file parallel read                                                    30 blocks
db file parallel read                                                    17 blocks
db file parallel read                                                    25 blocks
db file parallel read                                                    30 blocks
db file parallel read                                                    31 blocks
db file parallel read                                                    27 blocks
db file parallel read                                                    25 blocks
db file parallel read                                                    21 blocks
db file parallel read                                                    29 blocks
db file parallel read                                                    28 blocks

sudo -u root /usr/sbin/dtrace -q -n 'pid$target::pread:entry {_at_times["bytes"] = quantize(arg2);}' -p 10020

  bytes

           value  ------------- Distribution ------------- count
            4096 |                                         0
            8192 |_at_@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ 11215
           16384 |                                         0

As we can see above, the events are for the reads of 20-30 blocks. Each read is 8k. Arithmetically, the average wait time of 18ms is feasible for 20-30 8k reads, each of which taking a little bit less than 1ms in average.

> P1 does not match up to the file#, it is the number of files.

I completely misunderstood the meaning of P1. Thanks for pointing this out.

> I don’t think you can get the individual reads requests from a parallel read from the standard wait interface. There may be an additional event you can set or you might need to use OS tracing to see what files are being accessed.

I can easily find this out with DTrace if it will lead somewhere.

As a first aid, I could add more parallel slaves and/or increase the buffer cache.

A general question: why is recovery generating so many reads?

Best regards,

Nenad

https://nenadnoveljic.com/blog/



Please consider the environment before printing this e-mail. Bitte denken Sie an die Umwelt, bevor Sie dieses E-Mail drucken.

<html xmlns="http://www.w3.org/1999/xhtml">
<head>
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
<style type="text/css">p { font-family: Arial;font-size:9pt }</style>
</head>
<body>
<p>
<br>Important Notice</br>
<br />
This message is intended only for the individual named. It may contain confidential or privileged information. If you are not the named addressee you should in particular not disseminate, distribute, modify or copy this e-mail. Please notify the sender immediately by e-mail, if you have received this message by mistake and delete it from your system.<br /> Without prejudice to any contractual agreements between you and us which shall prevail in any case, we take it as your authorization to correspond with you by e-mail if you send us messages by e-mail. However, we reserve the right not to execute orders and instructions transmitted by e-mail at any time and without further explanation.<br /> E-mail transmission may not be secure or error-free as information could be intercepted, corrupted, lost, destroyed, arrive late or incomplete. Also processing of incoming e-mails cannot be guaranteed. All liability of Vontobel Holding Ltd. and any of its affiliates (hereinafter collectively referred to as "Vontobel Group") for any damages resulting from e-mail use is excluded. You are advised that urgent and time sensitive messages should not be sent by e-mail and if verification is required please request a printed version.</br> Please note that all e-mail communications to and from the Vontobel Group are subject to electronic storage and review by Vontobel Group. Unless stated to the contrary and without prejudice to any contractual agreements between you and Vontobel Group which shall prevail in any case, e-mail-communication is for informational purposes only and is not intended as an offer or solicitation for the purchase or sale of any financial instrument or as an official confirmation of any transaction.<br /> The legal basis for the processing of your personal data is the legitimate interest to develop a commercial relationship with you, as well as your consent to forward you commercial communications. You can exercise, at any time and under the terms established under current regulation, your rights. If you prefer not to receive any further communications, please contact your client relationship manager if you are a client of Vontobel Group or notify the sender. Please note for an exact reference to the affected group entity the corporate e-mail signature. For further information about data privacy at Vontobel Group please consult <a href="https://www.vontobel.com">www.vontobel.com</a>.<br />
</p>
</body>
</html>

--
http://www.freelists.org/webpage/oracle-l
Received on Thu Feb 27 2020 - 00:49:38 CET

Original text of this message