On Thu, Sep 24, 2020 at 3:32 AM Martin Braun
wrote:
> Hello all,
>
> i digged a bit deeper into this:
>
> the WALPlayer is not able to replay recovered.edits files, the source code
> http://hbase.apache.org/2.2/devapidocs/src-html/org/apache/hadoop/hbase/mapreduce/WALInputFormat.html
>
> seems
Let me check back in my notes for what step we're missing.
On Thu, Sep 24, 2020 at 5:32 AM Martin Braun wrote:
>
> Hello all,
>
> i digged a bit deeper into this:
>
> the WALPlayer is not able to replay recovered.edits files, the source code
>
Hello all,
i digged a bit deeper into this:
the WALPlayer is not able to replay recovered.edits files, the source code
http://hbase.apache.org/2.2/devapidocs/src-html/org/apache/hadoop/hbase/mapreduce/WALInputFormat.html
seems to expect an endtime coded into the filename:
long
Hello,
now I found some logs (see my previous post) they were directly written to
yarn.log file and not what I assumed to the map and reduce job log.
The WALPlayer claims that the recovered.edits are not WAL Files:
2020-09-23 12:43:09,061 WARN
Hello Sean,
thanks, as a map reduce job I was able to run the WALPlayer like this:
yarn jar "hbase-2.2.5/lib/shaded-clients/hbase-shaded-mapreduce-2.2.5.jar"
WALPlayer
"/hbase/data/default/tt_ix_parent_item/60de2aa0715e35657f5dd0e67bf07de8/recovered.edits/"
"tt_ix_parent_item"
However the
hurm. following the instructions from the reference guide works for
me. Is there a specific reason you're passing the
'--internal-classpath' flag? Do other hadoop jobs work?
what if you submit it as a proper MR job? unfortunately the ref guide
is thin on explaining this atm, but it looks like:
Hello Sean,
thank you for you quick response!
Replaying the wal files would be OK- however I am struggling using the
WALPlayer:
hbase --internal-classpath org.apache.hadoop.hbase.mapreduce.WALPlayer
bulk loading stuff works with hfiles. recovered.edits files are
formatted the same as WAL files rather than as HFiles. for wal files
you can use the wal replayer to ensure those edits are all present in
the table.
IIRC there is an unknown sequence of events that can result in the
recovered edits
Hello all,
I have an issue with hbase 2.2.5 (and hadoop-2.8.5) after a full disk event I
have 38 inconsistencies, when I do a
hbase --internal-classpath hbck
I get a bunch of these errors:
ERROR: Orphan region in HDFS: Unable to load .regioninfo from table
tt_ix_bizStep_inserting in hdfs
