Checkpoints: Storing before-images of changed pages - Revision history

Goff: Add tables X and E

2025-09-02T04:58:13Z

Add tables X and E

← Older revision		Revision as of 04:58, 2 September 2025
Line 511:		Line 511:

	<tr>		<tr>
	<td align="right">00</td>		<td align="right">x'00'</td>
	<td>File Control Table (FCT)</td>		<td>File Control Table (FCT)</td>
	</tr>		</tr>

	<tr>		<tr>
	<td align="right">01</td>		<td align="right">x'01'</td>
	<td>Table A</td>		<td>Table A</td>
	</tr>		</tr>

	<tr>		<tr>
	<td align="right">02</td>		<td align="right">x'02'</td>
	<td>Table B</td>		<td>Table B</td>
	</tr>		</tr>

	<tr>		<tr>
	<td align="right">03</td>		<td align="right">x'03'</td>
	<td>Table C</td>		<td>Table C</td>
	</tr>		</tr>

	<tr>		<tr>
	<td align="right">04</td>		<td align="right">x'04'</td>
	<td>Table D</td>		<td>Table D</td>
			</tr>

			<tr>
			<td align="right">x'2p'</td>
			<td>Table X
			Where <code>p</code> is the high order nibble of the page number which may range from 0 to 1 inclusive.
			</td>
			</tr>

			<tr>
			<td align="right">x'4p'</td>
			<td>Table E
			Where <code>p</code> is the high order nibble of the page number which may range from 0 to 3 inclusive.
			</td>
	</tr>		</tr>
	</table>		</table>

	</li>		</li>

	<li><code>PPPPPP</code> is a 3-byte hexadecimal table-relative page number.</li>		<li><code>PPPPPP</code> is a 3-byte hexadecimal table-relative page number.</li>
	</ul>		</ul>

JAL: /* Using sub-transaction checkpoints in recovery */ add link

2016-12-29T21:02:08Z

Using sub-transaction checkpoints in recovery: add link

← Older revision		Revision as of 21:02, 29 December 2016
Line 263:		Line 263:
	A sub-transaction checkpoint causes the current stream to be switched and checkpoint information written to the beginning of the new stream. If the Online comes down before the sub-transaction checkpoint completes then it is unusable and the previous good checkpoint must be used for restart recovery. </p>		A sub-transaction checkpoint causes the current stream to be switched and checkpoint information written to the beginning of the new stream. If the Online comes down before the sub-transaction checkpoint completes then it is unusable and the previous good checkpoint must be used for restart recovery. </p>
	<p>		<p>
	CCATEMP, <var>[[NDIR parameter\|NDIR]]</var>, and <var>[[NFILES parameter\|NFILES]]</var> must be the same size or greater than the settings used in the run being recovered.</p>		[[CCATEMP]], <var>[[NDIR parameter\|NDIR]]</var>, and <var>[[NFILES parameter\|NFILES]]</var> must be the same size or greater than the settings used in the run being recovered.</p>

	===Using transaction or sub-transaction checkpoints===		===Using transaction or sub-transaction checkpoints===
	<p>		<p>

DCameron: /* Example */

2016-12-20T16:37:22Z

Example

← Older revision		Revision as of 16:37, 20 December 2016
Line 197:		Line 197:
	ST X13,SAVEAREA+4 CHAIN CALLERS SAVEAREA TO OURS		ST X13,SAVEAREA+4 CHAIN CALLERS SAVEAREA TO OURS
	LA X13,SAVEAREA SET A(OUR SAVEAREA)		LA X13,SAVEAREA SET A(OUR SAVEAREA)
	WTO 'M204CKPX CHECKPOINT EXIT INVOKED,		WTO 'M204CKPX CHECKPOINT EXIT INVOKED, UPDATERS SUSPENDED'
	UPDATERS SUSPENDED'

	<b></b>* ********************************************************** *		<b></b>* ********************************************************** *
Line 229:		Line 228:
	<li>Left interrupts disabled, or did not restore the ESTAE or ESPIE macro routines. </li>		<li>Left interrupts disabled, or did not restore the ESTAE or ESPIE macro routines. </li>
	</ul></blockquote>		</ul></blockquote>

	==Overview of sub-transaction checkpoints==		==Overview of sub-transaction checkpoints==
	<p>		<p>

ELowell: /* Maintaining single volume CHKPOINT data sets */

2016-08-11T20:11:22Z

Maintaining single volume CHKPOINT data sets

← Older revision		Revision as of 20:11, 11 August 2016
Line 379:		Line 379:
	===Maintaining single volume CHKPOINT data sets===		===Maintaining single volume CHKPOINT data sets===
	<p>		<p>
	~~Multivolume~~ CHKPOINT data sets are not supported for roll back recovery because BSAM does not support read-backward for those kinds of data sets.		In Model 204 version 7.6 and earlier, multivolume CHKPOINT data sets are not supported for roll back recovery because BSAM does not support read-backward for those kinds of data sets. If a CHKPOINT data set is written to multiple volumes, then it is necessary to copy the data set onto a single volume before running <var>RESTART ROLL BACK</var> recovery.</p>
	~~</p>~~		<p>In Model 204 version 7.7 and later, multivolume CHKPOINT data sets are supported for roll back recovery because roll back pass 1 and pass 2 both read the recovery data set forward. See [[System and media recovery#ROLL BACK facility\|ROLL BACK facility]] for details.</p>
	~~<p>~~
	If a CHKPOINT data set is written to multiple volumes, then it is necessary to copy the data set onto a single volume before running <var>RESTART ROLL BACK</var> recovery.</p>
	<p>		<p>
	Large CHKPOINT data sets are usually not required when CPMAX=1 is set in CCAIN. This is explained in [[#Limiting the size of CHKPOINT\|Limiting the size of CHKPOINT]], below. GDGs (described in [[Configuring checkpoint and journal data streams]]) can be used for the CHKPOINT data set when long-running-update transactions are common.</p>		Large CHKPOINT data sets are usually not required when CPMAX=1 is set in CCAIN. This is explained in [[#Limiting the size of CHKPOINT\|Limiting the size of CHKPOINT]], below. GDGs (described in [[Configuring checkpoint and journal data streams]]) can be used for the CHKPOINT data set when long-running-update transactions are common.</p>

ELowell: /* Calculating disk space required */

2016-08-11T20:02:34Z

Calculating disk space required

← Older revision		Revision as of 20:02, 11 August 2016
Line 373:		Line 373:
	Calculate the maximum amount of disk space in use in this run in the CHKPOINT data set by dividing the high water mark of records (page preimages) written to the CHKPOINT data set by the number of pages per track for your device type. To determine pages per track, see [[File size calculation in detail#Data set allocation\|Data set allocation]]. </p>		Calculate the maximum amount of disk space in use in this run in the CHKPOINT data set by dividing the high water mark of records (page preimages) written to the CHKPOINT data set by the number of pages per track for your device type. To determine pages per track, see [[File size calculation in detail#Data set allocation\|Data set allocation]]. </p>
	<p>		<p>
	You can also monitor the size of the CHKPOINT data set using any existing utility appropriate for your operating system.</p>		You can also monitor the size of the CHKPOINT data set using any existing utility appropriate for your operating system. <var>[[MONITOR SIZE command\|MONITOR SIZE]] CHKP</var> is a useful command for this purpose.</p>
	<p>		<p>
	Generous sizing of the CHKPOINT data set is important, because <var class="product">Model 204</var> terminates whenever the CHKPOINT data set becomes full. For this reason, make the primary extent at least 50% greater than that calculated above. Further, define the data set with a substantial secondary extent, say 25% of the primary extent, and 15 extents of this size should be guaranteed available. For information about configuring the CHKPOINT data set, see [[Configuring checkpoint and journal data streams]].</p>		Generous sizing of the CHKPOINT data set is important, because <var class="product">Model 204</var> terminates whenever the CHKPOINT data set becomes full. For this reason, make the primary extent at least 50% greater than that calculated above. Further, define the data set with a substantial secondary extent, say 25% of the primary extent, and 15 extents of this size should be guaranteed available. For information about configuring the CHKPOINT data set, see [[Configuring checkpoint and journal data streams]].</p>

	===Maintaining single volume CHKPOINT data sets===		===Maintaining single volume CHKPOINT data sets===
	<p>		<p>

ELowell: /* Maintaining single volume CHKPOINT data sets */

2016-07-18T20:33:06Z

Maintaining single volume CHKPOINT data sets

← Older revision		Revision as of 20:33, 18 July 2016
Line 384:		Line 384:
	If a CHKPOINT data set is written to multiple volumes, then it is necessary to copy the data set onto a single volume before running <var>RESTART ROLL BACK</var> recovery.</p>		If a CHKPOINT data set is written to multiple volumes, then it is necessary to copy the data set onto a single volume before running <var>RESTART ROLL BACK</var> recovery.</p>
	<p>		<p>
	Large CHKPOINT data sets are usually not required when CPMAX=1 is set in CCAIN. This is explained in [[#Limiting the size of CHKPOINT\|Limiting the size of CHKPOINT]]. GDGs (described in [[Configuring checkpoint and journal data streams]]) can be used for the CHKPOINT data set when long-running-update transactions are common.</p>		Large CHKPOINT data sets are usually not required when CPMAX=1 is set in CCAIN. This is explained in [[#Limiting the size of CHKPOINT\|Limiting the size of CHKPOINT]], below. GDGs (described in [[Configuring checkpoint and journal data streams]]) can be used for the CHKPOINT data set when long-running-update transactions are common.</p>

	<p>If checkpoint timeouts (which can require large CHKPOINT data sets) are common, the use of [[#Overview of sub-transaction checkpoints\|sub-transaction checkpoints]] is recommended. Sub-transaction checkpoints eliminate checkpoint timeouts and help to keep the size of the CHKPOINT data set to a minimum.		<p>If checkpoint timeouts (which can require large CHKPOINT data sets) are common, the use of [[#Overview of sub-transaction checkpoints\|sub-transaction checkpoints]] is recommended. Sub-transaction checkpoints eliminate checkpoint timeouts and help to keep the size of the CHKPOINT data set to a minimum.

ELowell: /* For more information */

2016-07-18T20:27:02Z

For more information

← Older revision		Revision as of 20:27, 18 July 2016
Line 21:		Line 21:
	This page discusses the Checkpoint facility, preimage logging, and dumping the CHKPOINT file with the UTILC utility. </p>		This page discusses the Checkpoint facility, preimage logging, and dumping the CHKPOINT file with the UTILC utility. </p>
	<ul>		<ul>
	<li>HLI ~~preimage logging is~~ discussed in detail in ~~the <var class="book">~~[[~~Media~~:~~M204 HLIReference V75.pdf\|Rocket~~ Model 204 ~~Host Language Interface Reference Manual~~]].~~</var>~~</li>		<li>Checkpoints in HLI are discussed in detail in [[HLI: Model 204 recovery and checkpoints]].</li>

	<li>[[System and media recovery#Recovery data sets and job control\|Recovery data sets and job control]] explains how to use CHKPOINT in the recovery process.</li>		<li>[[System and media recovery#Recovery data sets and job control\|Recovery data sets and job control]] explains how to use CHKPOINT in the recovery process.</li>
Line 27:		Line 27:
	<li>Checkpoint data stream configurations are described in [[Configuring checkpoint and journal data streams]]. </li>		<li>Checkpoint data stream configurations are described in [[Configuring checkpoint and journal data streams]]. </li>
	</ul>		</ul>

	==Understanding the Checkpoint facility==		==Understanding the Checkpoint facility==
	<p>		<p>

ELowell: /* Maintaining single volume CHKPOINT data sets */

2016-03-31T20:38:42Z

Maintaining single volume CHKPOINT data sets

← Older revision		Revision as of 20:38, 31 March 2016
Line 384:		Line 384:
	If a CHKPOINT data set is written to multiple volumes, then it is necessary to copy the data set onto a single volume before running <var>RESTART ROLL BACK</var> recovery.</p>		If a CHKPOINT data set is written to multiple volumes, then it is necessary to copy the data set onto a single volume before running <var>RESTART ROLL BACK</var> recovery.</p>
	<p>		<p>
	Large CHKPOINT data sets are usually not required when CPMAX=1 is set in CCAIN. This is explained in [[#Limiting the size of CHKPOINT\|Limiting the size of CHKPOINT]]. GDGs ([[~~#Configuring checkpoint and journal data streams\|~~Configuring checkpoint and journal data streams]]) can be used for the CHKPOINT data set when long-running-update transactions are common. If checkpoint timeouts (which can require large CHKPOINT data sets) are common, the use of [[#Overview of sub-transaction checkpoints\|sub-transaction checkpoints]] is recommended. Sub-transaction checkpoints eliminate checkpoint timeouts and help to keep the size of the CHKPOINT data set to a minimum.		Large CHKPOINT data sets are usually not required when CPMAX=1 is set in CCAIN. This is explained in [[#Limiting the size of CHKPOINT\|Limiting the size of CHKPOINT]]. GDGs (described in [[Configuring checkpoint and journal data streams]]) can be used for the CHKPOINT data set when long-running-update transactions are common.</p>

			<p>If checkpoint timeouts (which can require large CHKPOINT data sets) are common, the use of [[#Overview of sub-transaction checkpoints\|sub-transaction checkpoints]] is recommended. Sub-transaction checkpoints eliminate checkpoint timeouts and help to keep the size of the CHKPOINT data set to a minimum.
	</p>		</p>

ELowell: /* Maintaining single volume CHKPOINT data sets */

2016-03-31T20:23:24Z

Maintaining single volume CHKPOINT data sets

← Older revision		Revision as of 20:23, 31 March 2016
Line 384:		Line 384:
	If a CHKPOINT data set is written to multiple volumes, then it is necessary to copy the data set onto a single volume before running <var>RESTART ROLL BACK</var> recovery.</p>		If a CHKPOINT data set is written to multiple volumes, then it is necessary to copy the data set onto a single volume before running <var>RESTART ROLL BACK</var> recovery.</p>
	<p>		<p>
	Large CHKPOINT data sets are usually not required when CPMAX=1 is set in CCAIN. This is explained in [[#Limiting the size of CHKPOINT\|Limiting the size of CHKPOINT]]. GDGs ([[#Configuring checkpoint and journal data streams\|Configuring checkpoint and journal data streams]]) can be used for the CHKPOINT data set when long-running-update transactions are common. If checkpoint timeouts, which can require large CHKPOINT data sets are common, the use of ~~sub-transaction checkpoints (~~[[#Overview of sub-transaction checkpoints\|~~Overview of~~ sub-transaction checkpoints]]) is recommended~~; sub~~-transaction checkpoints eliminate checkpoint timeouts and help to keep the size of the CHKPOINT data set to a minimum.		Large CHKPOINT data sets are usually not required when CPMAX=1 is set in CCAIN. This is explained in [[#Limiting the size of CHKPOINT\|Limiting the size of CHKPOINT]]. GDGs ([[#Configuring checkpoint and journal data streams\|Configuring checkpoint and journal data streams]]) can be used for the CHKPOINT data set when long-running-update transactions are common. If checkpoint timeouts (which can require large CHKPOINT data sets) are common, the use of [[#Overview of sub-transaction checkpoints\|sub-transaction checkpoints]] is recommended. Sub-transaction checkpoints eliminate checkpoint timeouts and help to keep the size of the CHKPOINT data set to a minimum.
	</p>		</p>

JDamon: Rewrote this section since it was confusing and misleading

2016-03-30T00:01:15Z

Rewrote this section since it was confusing and misleading

← Older revision		Revision as of 00:01, 30 March 2016
Line 379:		Line 379:
	===Maintaining single volume CHKPOINT data sets===		===Maintaining single volume CHKPOINT data sets===
	<p>		<p>
	~~You cannot process multivolume~~ CHKPOINT data sets ~~during~~ recovery because of ~~limitations of the BSAM access method used by <var>RESTART</var> recovery~~.</p>		Multivolume CHKPOINT data sets are not supported for roll back recovery because BSAM does not support read-backward for those kinds of data sets.
			</p>
	<p>		<p>
	If a CHKPOINT data set is written to multiple volumes~~, because of secondary allocation under z/OS~~, then it is necessary to copy ~~this file~~ onto a single volume before running <var>RESTART</var> recovery.</p>		If a CHKPOINT data set is written to multiple volumes, then it is necessary to copy the data set onto a single volume before running <var>RESTART ROLL BACK</var> recovery.</p>
	<p>		<p>
	~~Avoid multivolume~~ CHKPOINT data sets ~~by making sure that primary disk allocation~~ is ~~adequate, or by setting the <var>CPMAX</var> parameter (~~explained ~~as follows~~ in [[#Limiting the size of CHKPOINT\|Limiting the size of CHKPOINT]]) to a ~~sufficiently low value~~.</p>		Large CHKPOINT data sets are usually not required when CPMAX=1 is set in CCAIN. This is explained in [[#Limiting the size of CHKPOINT\|Limiting the size of CHKPOINT]]. GDGs ([[#Configuring checkpoint and journal data streams\|Configuring checkpoint and journal data streams]]) can be used for the CHKPOINT data set when long-running-update transactions are common. If checkpoint timeouts, which can require large CHKPOINT data sets are common, the use of sub-transaction checkpoints ([[#Overview of sub-transaction checkpoints\|Overview of sub-transaction checkpoints]]) is recommended; sub-transaction checkpoints eliminate checkpoint timeouts and help to keep the size of the CHKPOINT data set to a minimum.
			</p>

	===Limiting the size of CHKPOINT===		===Limiting the size of CHKPOINT===
	<p>		<p>