[ https://issues.apache.org/jira/browse/HBASE-9915?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13816350#comment-13816350 ]
Lars Hofhansl commented on HBASE-9915: -------------------------------------- [~jmspaggi] Yeah, did some count queries in Phoenix. Without patch they took 27s with the patch they take 14s. You'll see the improvement if (1) you use a block encoder (like FAST_DIFF, etc) and (2) you added some columns to you scan object (so that the ExplicitColumnTracker is used under the hood). I am not sure any the the performance evaluation tests do that (and if not, we should probably add that). Making a trunk patch now for a full test run. > Severe performance bug: isSeeked() in EncodedScannerV2 is always false > ---------------------------------------------------------------------- > > Key: HBASE-9915 > URL: https://issues.apache.org/jira/browse/HBASE-9915 > Project: HBase > Issue Type: Bug > Reporter: Lars Hofhansl > Assignee: Lars Hofhansl > Fix For: 0.98.0, 0.96.1, 0.94.14 > > Attachments: 9915-0.94-v2.txt, 9915-0.94.txt, profile.png > > > While debugging why reseek is so slow I found that it is quite broken for > encoded scanners. > The problem is this: > AbstractScannerV2.reseekTo(...) calls isSeeked() to check whether scanner was > seeked or not. If it was it checks whether the KV we want to seek to is in > the current block, if not it always consults the index blocks again. > isSeeked checks the blockBuffer member, which is not used by EncodedScannerV2 > and thus always returns false, which in turns causes an index lookup for each > reseek. -- This message was sent by Atlassian JIRA (v6.1#6144)