The CR blocks created: Oracle metric is the number of CURRENT blocks cloned to create CR (consistent read) blocks. The most common reason for cloning is that the buffer is held in an incompatible mode.
Read consistency is used when competing processes are reading and updating the data concurrently.
The gc cr block receive time:Oracle metric isGlobal cache cr block receive timeStatistic records the total time required for consistent read requests to complete. In other words, it records the accumulated round-trip time for all requests for consistent read blocks.
When we want to search for a data block, we find that the version of this data block is newer than what we want to search.UNDOTo find the pre-image of the data (PRE IMAGE), Find the pre-image of the data in the rollback segmentCURRENTData block merging,To formCR BLOCKIn this way,QueryCR BLOCKYou can get consistent data..
ThisCR BLOCKIs not generated inPGAAndYesDB CACHEApply for a data block, and then generateCR BLOCK. ThisCR BLOCKAfter creation, if you want to access the data block of this version next time, you do not need to create a newCR BLOCKYou can directly use the previously createdCR BLOCK.Each createdCR BLOCK,CR BLOCKS CREATEDThe counter is added.1.
BecauseCR BLOCKAndCURRENT BLOCKOfRDBAThey are all the same, so they will be placed in the sameHASHChain.Therefore, if someBLOCKOfCR BLOCKToo many versions will also causeBUFFER CACHE CHAINSCompetition for latches is intensifying.
RelatedRDBAFor more information, see myBlog:
Oracle rdbaAndDbaDescription
IfCR BLOCKIn the process of discoveryUNDOInPRE IMAGEBecause the submission time is long and has been overwritten, the famousORA-1555.
IfCR BLOCKS CREATEDLarge, then you should check the system'sCPUCheck whether the resource has a bottleneck.BUFFER CACHEWhether there is competition for related latches.
Buffer Cache ChianCause:
When a session needs to access a memory block, it first needs to search for the block in the memory in a structure like a linked list, when a session accesses this linked list, you need to obtainLatchIf the query failsLatch cache buffer chainWait,The reason for this wait is that there are too many sessions accessing the same data block or the list is too long (if too much data is read in the memory,TheHashThe list will be very long, so that the time for the session scan list will increase.Chache buffer chain latchThe time will get longer, and other sessions will get thisLatchThe opportunity will decrease, and the wait will increase ).
RelatedLatchFor more information, seeBlog:
Oracle LatchDescription