<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"
	xmlns:content="http://purl.org/rss/1.0/modules/content/"
	xmlns:dc="http://purl.org/dc/elements/1.1/"
	xmlns:atom="http://www.w3.org/2005/Atom"
	xmlns:sy="http://purl.org/rss/1.0/modules/syndication/"
		>
<channel>
	<title>Comments on: Diagnostics For Database Hang</title>
	<atom:link href="http://askdba.org/weblog/2008/03/diagnostics-for-database-hang/feed/" rel="self" type="application/rss+xml" />
	<link>http://askdba.org/weblog/2008/03/diagnostics-for-database-hang/</link>
	<description>Writing About Our Experiences With Oracle Databases</description>
	<lastBuildDate>Fri, 05 Mar 2010 10:24:30 +0000</lastBuildDate>
	<generator>http://wordpress.org/?v=2.9.2</generator>
	<sy:updatePeriod>hourly</sy:updatePeriod>
	<sy:updateFrequency>1</sy:updateFrequency>
		<item>
		<title>By: Amit</title>
		<link>http://askdba.org/weblog/2008/03/diagnostics-for-database-hang/comment-page-1/#comment-113</link>
		<dc:creator>Amit</dc:creator>
		<pubDate>Thu, 07 Aug 2008 03:53:25 +0000</pubDate>
		<guid isPermaLink="false">http://askoracledba.wordpress.com/2008/03/21/diagnostics-for-database-hang/#comment-113</guid>
		<description>You need to first check database alert log for any errors/issues. Sometimes errors like ORA-600 and ORA-7445 are reported which could be causing the hang. Also when hangs are caused by latching or enqueue issues such as &quot;PMON failed to acquire latch&quot;, then you will see that systemstate will be dumped automatically. This can be uploaded to metalink.

Also use OS utilities like top,vmstat if there is CPU contention and OS logs (/var/log/messages on linux).
In case your Database is hung in such a way that you cannot run queries, then you need to take Systemstate dump. You can use below note for same in 8i/9i

Note 121779.1-Taking a SYSTEMSTATE dump when you cannot CONNECT to Oracle.

In case of 10g, you can use prelim to connect to database.
e.g
sqlplus -prelim “/ as sysdba” and then take systemstate and hanganalyze

-Amit</description>
		<content:encoded><![CDATA[<p>You need to first check database alert log for any errors/issues. Sometimes errors like ORA-600 and ORA-7445 are reported which could be causing the hang. Also when hangs are caused by latching or enqueue issues such as &#8220;PMON failed to acquire latch&#8221;, then you will see that systemstate will be dumped automatically. This can be uploaded to metalink.</p>
<p>Also use OS utilities like top,vmstat if there is CPU contention and OS logs (/var/log/messages on linux).<br />
In case your Database is hung in such a way that you cannot run queries, then you need to take Systemstate dump. You can use below note for same in 8i/9i</p>
<p>Note 121779.1-Taking a SYSTEMSTATE dump when you cannot CONNECT to Oracle.</p>
<p>In case of 10g, you can use prelim to connect to database.<br />
e.g<br />
sqlplus -prelim “/ as sysdba” and then take systemstate and hanganalyze</p>
<p>-Amit</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: siva</title>
		<link>http://askdba.org/weblog/2008/03/diagnostics-for-database-hang/comment-page-1/#comment-112</link>
		<dc:creator>siva</dc:creator>
		<pubDate>Thu, 07 Aug 2008 03:07:17 +0000</pubDate>
		<guid isPermaLink="false">http://askoracledba.wordpress.com/2008/03/21/diagnostics-for-database-hang/#comment-112</guid>
		<description>Very good information on database hang.
But i have few questions on this:

If database  hung,we will not generate any reports,not running queries also.In this situation,what is the first step we ill take?</description>
		<content:encoded><![CDATA[<p>Very good information on database hang.<br />
But i have few questions on this:</p>
<p>If database  hung,we will not generate any reports,not running queries also.In this situation,what is the first step we ill take?</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Amit</title>
		<link>http://askdba.org/weblog/2008/03/diagnostics-for-database-hang/comment-page-1/#comment-20</link>
		<dc:creator>Amit</dc:creator>
		<pubDate>Sat, 05 Apr 2008 01:28:00 +0000</pubDate>
		<guid isPermaLink="false">http://askoracledba.wordpress.com/2008/03/21/diagnostics-for-database-hang/#comment-20</guid>
		<description>Aarya,You can use P1,P2,P3 values to diagnose the parameters for wait event. e.g You see event &#039;latch free&#039; as considerable wait event, then
P1 = Latch address
P2 = Latch number

Now you can identify the latch name from V$LATCHNAME
SELECT * FROM v$latchname  WHERE latch# =P2;
Moreover there can be situation where many sessions are waiting on same latch anf their P1 and P2 values are same say it  points to Shared Pool Latch. From this we can be sure that some latching issue is causing the Database Hang.I will try to take up a case study sometime so that it can be more clear.And for your second question, you can refer to
Note 121779.1-Taking a SYSTEMSTATE dump when you cannot CONNECT to Oracle.

 It is already mentioned in the post.In case of 10g, you can use &lt;i&gt;prelim &lt;/i&gt; to connect to database.
e.g
sqlplus -prelim &quot;/ as sysdba&quot;

Cheers
Amit</description>
		<content:encoded><![CDATA[<p>Aarya,You can use P1,P2,P3 values to diagnose the parameters for wait event. e.g You see event &#8216;latch free&#8217; as considerable wait event, then<br />
P1 = Latch address<br />
P2 = Latch number</p>
<p>Now you can identify the latch name from V$LATCHNAME<br />
SELECT * FROM v$latchname  WHERE latch# =P2;<br />
Moreover there can be situation where many sessions are waiting on same latch anf their P1 and P2 values are same say it  points to Shared Pool Latch. From this we can be sure that some latching issue is causing the Database Hang.I will try to take up a case study sometime so that it can be more clear.And for your second question, you can refer to<br />
Note 121779.1-Taking a SYSTEMSTATE dump when you cannot CONNECT to Oracle.</p>
<p> It is already mentioned in the post.In case of 10g, you can use <i>prelim </i> to connect to database.<br />
e.g<br />
sqlplus -prelim &#8220;/ as sysdba&#8221;</p>
<p>Cheers<br />
Amit</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Aaryadba</title>
		<link>http://askdba.org/weblog/2008/03/diagnostics-for-database-hang/comment-page-1/#comment-19</link>
		<dc:creator>Aaryadba</dc:creator>
		<pubDate>Fri, 21 Mar 2008 14:36:00 +0000</pubDate>
		<guid isPermaLink="false">http://askoracledba.wordpress.com/2008/03/21/diagnostics-for-database-hang/#comment-19</guid>
		<description>Hi Piyush,&lt;br/&gt;&lt;br/&gt;Very good information on database hang.&lt;br/&gt;But i have few questions on this:&lt;br/&gt;&lt;br/&gt;-What does p1,p2,p3 represents in  v$session_wait and what are their significanse here.&lt;br/&gt;- I have faced this situation many a times but i was not able to login to database when the hang occured, so what to do in that case, i.e how to generate dumps.</description>
		<content:encoded><![CDATA[<p>Hi Piyush,</p>
<p>Very good information on database hang.<br />But i have few questions on this:</p>
<p>-What does p1,p2,p3 represents in  v$session_wait and what are their significanse here.<br />- I have faced this situation many a times but i was not able to login to database when the hang occured, so what to do in that case, i.e how to generate dumps.</p>
]]></content:encoded>
	</item>
</channel>
</rss>
