hbase-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jean-Marc Spaggiari <jean-m...@spaggiari.org>
Subject Re: [VOTE] The 1st hbase 0.94.11 release candidate is available for download
Date Tue, 20 Aug 2013 13:52:28 GMT
So.
I have downloaded all the files and verified the signature for security and
regular version. Uncompressed both of them, and checked the CHANGES.txt
file. All fine.

Looked at the documentation on random pages, still fine.

I ran the test suite successfuly. Ran standalone, checked the logs and UI,
fine again.

Performance Evaluation is still running and will most probably take another
24h, so no results yet. I will post them when ready. I don't think there is
any impact there.
Deployed the jars on a 9 nodes cluster (8 RS). Did rolling restart
--graceful -maxthreads 30, still fine.
Checked the logs, the web ui, etc. Fine.
Change the load balancer to default, stopped the cluster, merged a table to
a single region table, restart the cluster, major_compact. Table got
splitted correctly, and balancer re-balanced the regions almost correctly.
Stopped the cluser, re-merged, restored customed balancer, restart,
re-compact, etc. all fine. Regions got spliited, re-assigned,
transitionned, all fine.
Ran a MR job on the same cluster, still fine.


tl.tr.

1) Still many DEBUG lines in the logs (Not new HBASE-9037)
2013-08-20 09:49:39,552 DEBUG
org.apache.hadoop.hbase.io.hfile.LruBlockCache: Block cache LRU eviction
started; Attempting to free 24.84 MB of total=210.78 MB
2013-08-20 09:49:39,556 DEBUG
org.apache.hadoop.hbase.io.hfile.LruBlockCache: Block cache LRU eviction
completed; freed=24.85 MB, total=186.06 MB, single=114.14 MB, multi=94.6
MB, memory=5.12 KB

2) Balancer need to run twice to get the regions balanced correctly (Not
new)

+1 for me.

JM


2013/8/19 Jean-Marc Spaggiari <jean-marc@spaggiari.org>

> Actively working on that. PE is running in both 94.10 and 94.11 right now.
> Will have numbers by tomorrow morning. So give me another 12h ;)
>
> JM
> Le 2013-08-19 20:02, "lars hofhansl" <larsh@apache.org> a écrit :
>
> One additional +1 please.
>>
>> Release testing is about as exiting as getting a tooth pulled, but I
>> can't release without it :)
>>
>>
>> -- Lars
>>
>>
>>
>> ________________________________
>>  From: Jean-Daniel Cryans <jdcryans@apache.org>
>> To: "dev@hbase.apache.org" <dev@hbase.apache.org>; lars hofhansl <
>> larsh@apache.org>
>> Sent: Thursday, August 15, 2013 10:21 AM
>> Subject: Re: [VOTE] The 1st hbase 0.94.11 release candidate is available
>> for download
>>
>>
>>
>> +1
>>
>> Used both tars against each other, loaded some PE data, ran scans and
>> reads, checked out the logs/UIs, did some jstack'ing to make sure nothing
>> weird was in there.
>>
>> J-D
>>
>>
>>
>> On Tue, Aug 13, 2013 at 11:26 PM, lars hofhansl <larsh@apache.org> wrote:
>>
>> The 1st 0.94.11 RC is available for download at
>> http://people.apache.org/~larsh/hbase-0.94.11-rc0/
>> >Signed with my code signing key: C7CFE328
>> >
>> >As before 0.94.11 is mostly a bug fix release and has 39 issues resolved
>> against it:
>> >    [HBASE-8779] - Add mutateRow method support to Thrift2
>> >    [HBASE-8946] - Add a new function to Thrift 2 to open scanner, get
>> results and close scanner
>> >    [HBASE-8947] - Thrift 2 : Replace "bool writeToWAL" with
>> "TDurability durability"
>> >    [HBASE-8948] - Bound table pool size in Thrift 2 server
>> >    [HBASE-6826] - [WINDOWS] TestFromClientSide failures
>> >    [HBASE-8067] - TestHFileArchiving.testArchiveOnTableDelete sometimes
>> fails
>> >    [HBASE-8670] - [0.94] Backport HBASE-8449,HBASE-8204 and HBASE-8699
>> to 0.94 (Refactor recoverLease retries and pauses)
>> >    [HBASE-8698] - potential thread creation in MetaScanner.metaScan
>> >    [HBASE-8935] - IntegrationTestBigLinkedList fails under load on 0.94
>> due to some scan issues - add logging
>> >    [HBASE-8949] - hbase.mapreduce.hfileoutputformat.blocksize should
>> configure with blocksize of a table
>> >    [HBASE-9026] - RestartRsHoldingRoot action in
>> org.apache.hadoop.hbase.util.ChaosMonkey restarting the server holding
>> .META. instead of -ROOT-
>> >    [HBASE-9032] - Result.getBytes() returns null if backed by KeyValue
>> array
>> >    [HBASE-9048] - HCM throws NullPointerException under load
>> >    [HBASE-9050] - HBaseClient#call could hang
>> >    [HBASE-9060] - ExportSnapshot job fails if target path contains
>> percentage character
>> >    [HBASE-9079] - FilterList getNextKeyHint skips rows that should be
>> included in the results
>> >    [HBASE-9080] - Retain assignment should be used when re-enabling
>> table(s)
>> >    [HBASE-9085] - Integration Tests fails because of bug in teardown
>> phase where the cluster state is not being restored properly.
>> >    [HBASE-9087] - Handlers being blocked during reads
>> >    [HBASE-9097] - Set HBASE_CLASSPATH before rest of the classpath
>> >    [HBASE-9115] - HTableInterface.append operation may overwrites values
>> >    [HBASE-9120] - ClassFinder logs errors that are not
>> >    [HBASE-9146] - TestHTablePool hangs when run as part of
>> runMediumTests profile
>> >    [HBASE-9158] - Serious bug in cyclic replication
>> >    [HBASE-9189] - IncreasingToUpperBoundRegionSplitPolicy.shouldSplit()
>> should check all the stores before returning.
>> >    [HBASE-9200] - HFilePrettyPrinter finds incorrect largest row
>> >    [HBASE-6580] - Deprecate HTablePool in favor of
>> HConnection.getTable(...)
>> >    [HBASE-8995] - Add hadoop-1.2 profile
>> >    [HBASE-9019] - Port HBASE-8690: Reduce unnecessary getFileStatus
>> hdfs calls in TTL hfile and hlog cleanners to 0.94
>> >    [HBASE-9029] - Backport HBASE-8706 Some improvement in snapshot to
>> 0.94
>> >    [HBASE-9132] - Use table dir modtime to avoid scanning table dir to
>> check cached table descriptor in 0.94
>> >    [HBASE-9139] - Independent timeout configuration for rpc channel
>> between cluster nodes
>> >    [HBASE-7826] - Improve Hbase Thrift v1 to return results in sorted
>> order
>> >    [HBASE-8663] - a HBase Shell command to list the tables replicated
>> from current cluster
>> >    [HBASE-8816] - Add support of loading multiple tables into
>> LoadTestTool
>> >    [HBASE-9075] - [0.94] Backport HBASE-5760 Unit tests should write
>> only under /target to 0.94
>> >    [HBASE-9090] - cleanup snapshot tests setup/teardown code
>> >    [HBASE-9106] - Do not fail TestAcidGuarantees for exceptions on
>> table flush
>> >    [HBASE-9107] - [0.94] Backport HBASE-6950 TestAcidGuarantees system
>> test now flushes too aggressively to 0.94
>> >
>> >The list of changes is also available here:
>> https://issues.apache.org/jira/secure/ReleaseNote.jspa?projectId=12310753&version=12324741
>> >
>> >Here're the jenkins runs for this RC:
>> https://builds.apache.org/job/HBase-0.94.11/3/ and
>> https://builds.apache.org/job/HBase-0.94.11-security/2/
>> >
>> >Please try out the RC, check out the doc, take it for a spin, etc, and
>> vote +1/-1 by August 20th on whether we should release this as 0.94.11.
>> >
>> >Thanks.
>> >
>> >-- Lars
>> >
>> >ps. nobody used the release testing sheet for the past few 0.94 point
>> releases, so I did not create one this time.
>> >
>
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message