Bug 30159581 - alter new primary open hang after switchover (Doc ID 30159581.8)
Bug 30159581 alter new primary open hang after switchover This note gives a brief overview of bug 30159581. The content was last updated on: 20-OCT-2020 Click here for details of each of the sections below. Affects: Product (Component) Oracle Server (Rdbms) Range of versions believed to be affected Versions BELOW 20.1 Versions confirmed as being affected 19.8.0 19.7.0 19.6.0 19.5.0 Platforms affected Generic (all / most platforms affected) Note that this fix can cause / expose the problem described in Bug:31747989 Note that this fix has been superseded by the fix in Bug:31747989 Fixed: This fix has been superseded - please see the fixed version information for Bug:31747989 . The box below only shows versions where the code change/s for 30159581 are first included - those versions may not contain the later improved fix. The fix for 30159581 is first included in 20.1.0 Symptoms: Related To: Hang (Involving Shared Resource) Instance Startup
Description This bug manifests as a hang during the opening of a database in read-write mode or during the switchover of a standby instance to the role of a primary instance.
We speed up the opening of the database by using parallel slaves to perform first-pass transaction recovery.
This phase detects all in-flight transactions from the past and mark them as such for subsequent recovery.
The hang occurs due to a deadlock between two parallel recovery slaves on dictionary cache objects. Rediscovery Notes A hang during the opening of a database in read-write mode or during the switchover of a standby instance to the role of a primary instance.
The database needs to have enough undo segments in "in-use" or "partially-available" states to warrant parallelization of FPTR.
Typically this is 100 (controlled by parameter _min_undosegs_for_parallel_fptr. Note: This fix supersedes the fix in bug 28770119. Workaround Turn off parallelization of first-pass transaction recovery (occurs as part of database open) by setting _min_undosegs_for_parallel_fptr = 0
in the initialization file and open the database.
Please note: The above is a summary description only. Actual symptoms can vary. Matching to any symptoms here does not confirm that you are encountering this problem. For questions about this bug please consult Oracle Support. References Bug:30159581 (This link will only work for PUBLISHED bugs) Note:245840.1 Information on the sections in this article
=======================================================================================================================================================
Bug 31747989 - tracking bug for dependent fixes of first-pass transaction recovery (Doc ID 31747989.8)
Bug 31747989 tracking bug for dependent fixes of first-pass transaction recovery This note gives a brief overview of bug 31747989. The content was last updated on: 02-DEC-2020 Click here for details of each of the sections below. Affects: Product (Component) Oracle Server (Rdbms) Range of versions believed to be affected Versions BELOW 20 Versions confirmed as being affected (None Specified) Platforms affected Generic (all / most platforms affected) Fixed: The fix for 31747989 is first included in 19.9.0.0.DBRU:201020 (OCT 2020) DB Release Update Revision(DB RU) Interim patches may be available for earlier versions - click here to check. Symptoms: Related To: Deadlock Hang (Involving Shared Resource) Performance Affected (General) Instance Startup HOT19C Description Note: This fix supersedes the fix in bug 30159581. For interim patch please use this fix instead of the fix in the bug 30159581 to address both the issues. Please note: The above is a summary description only. Actual symptoms can vary. Matching to any symptoms here does not confirm that you are encountering this problem. For questions about this bug please consult Oracle Support. References Bug:31747989 (This link will only work for PUBLISHED bugs) Note:245840.1 Information on the sections in this article