Having a look at your master log file from your 0.20.5 clean run could probably uncover what went wrong. Can you put it in a pastebin?
Also I would guess that it works well with 0.20.3. J-D On Wed, Jul 21, 2010 at 12:19 PM, Karthik Kambatla <[email protected]> wrote: > I have tried 0.20.4 and it behaves the same way. Finally, I am back to > 0.20.1, and it works fine. > > Thanks > Karthik > > On Wed, Jul 21, 2010 at 10:43 AM, Karthik Kambatla > <[email protected]>wrote: > >> Thanks a lot for the prompt help. [I had sent an immediate reply but from >> the wrong address... hopefully, you would receive this.] >> >> >> I am running HBase 0.20.5. I have cleaned the whole cluster and re-setup >> everything (HDFS, MapReduce, and HBase). Here is the transcript -- from when >> the HBase cluster has been started. >> >> >> [kamba...@mercado-2 ~]$ hbase shell >> HBase Shell; enter 'help<RETURN>' for list of supported commands. >> Version: 0.20.5, rUnknown, Thu Jul 15 22:27:05 PDT 2010 >> >> hbase(main):001:0> list >> 0 row(s) in 0.0910 seconds >> >> hbase(main):002:0> create 'Test' >> 0 row(s) in 1.0990 seconds >> >> hbase(main):003:0> list >> Test >> >> >> 1 row(s) in 0.0070 seconds >> >> hbase(main):004:0> exit >> >> >> [kamba...@mercado-2 ~]$ hbase shell >> HBase Shell; enter 'help<RETURN>' for list of supported commands. >> Version: 0.20.5, rUnknown, Thu Jul 15 22:27:05 PDT 2010 >> >> hbase(main):001:0> list >> 0 row(s) in 0.0860 seconds >> >> hbase(main):004:0> scan '.META.' >> ROW >> COLUMN+CELL >> >> Test,,1279665697803 column=info:server, timestamp=1279668781099, >> value=15.25.119.96:60020 >> Test,,1279665697803 column=info:serverstartcode, >> timestamp=1279668781099, value=1279665651801 >> 1 row(s) in 0.0260 seconds >> >> hbase(main):005:0> create 'Test1' >> 0 row(s) in 1.1050 seconds >> >> hbase(main):006:0> list >> 0 row(s) in 0.0060 seconds >> >> hbase(main):007:0> scan '.META.' >> ROW >> COLUMN+CELL >> >> Test,,1279665697803 column=info:server, timestamp=1279668781099, >> value=15.25.119.96:60020 >> Test,,1279665697803 column=info:serverstartcode, >> timestamp=1279668781099, value=1279665651801 >> 1 row(s) in 0.0100 seconds >> >> hbase(main):008:0> exit >> >> >> [kamba...@mercado-2 ~]$ hbase shell >> HBase Shell; enter 'help<RETURN>' for list of supported commands. >> Version: 0.20.5, rUnknown, Thu Jul 15 22:27:05 PDT 2010 >> >> hbase(main):001:0> list >> 0 row(s) in 0.0900 seconds >> >> hbase(main):002:0> scan '.META.' >> ROW >> COLUMN+CELL >> >> 0 row(s) in 0.0130 seconds >> >> hbase(main):003:0> scan '.META.' >> ROW >> COLUMN+CELL >> >> 0 row(s) in 0.0070 seconds >> >> Thanks >> Karthik >> >> >> On Tue, Jul 20, 2010 at 4:15 PM, Jonathan Gray <[email protected]> wrote: >> >>> Also, I'm not sure this will be fixed with the master changes. >>> >>> tableExists() actually makes a call to listTables() and then checks if the >>> table you are looking for is in the list. >>> >>> listTables() uses a MetaScanner/MetaVisitor. Though the master changes >>> de-emphasize the general use of meta scanning, it would not necessarily >>> change the implementation of these client (to RS) calls. >>> >>> Hard to say what would make this flakey besides some of the older bugs >>> around lots of META StoreFiles. >>> >>> What version of HBase are you running? >>> >>> JG >>> >>> > -----Original Message----- >>> > From: Jonathan Gray [mailto:[email protected]] >>> > Sent: Tuesday, July 20, 2010 4:06 PM >>> > To: [email protected] >>> > Subject: RE: Flaky tableExists() >>> > >>> > I've personally never seen a flaky tableExists() but it's not something >>> > I've used heavily. >>> > >>> > Could you do the same operations you did below in the shell but also >>> > run: scan '.META.' >>> > >>> > > -----Original Message----- >>> > > From: Karthik Kambatla [mailto:[email protected]] >>> > > Sent: Tuesday, July 20, 2010 3:26 PM >>> > > To: [email protected] >>> > > Subject: Re: Flaky tableExists() >>> > > >>> > > Thanks Ryan for understanding the pain :) >>> > > >>> > > Is there a work-around for the time being? I haven't experienced any >>> > > such >>> > > issues with earlier versions of HBase. Was I just lucky or would it >>> > > make >>> > > sense to revert to earlier questions? >>> > > >>> > > Thanks >>> > > Karthik >>> > > >>> > > On Tue, Jul 20, 2010 at 2:54 PM, Ryan Rawson <[email protected]> >>> > > wrote: >>> > > >>> > > > This sucks, and there are being substantial reworks to master >>> > > > functions in 0.90 :-) >>> > > > >>> > > > Hopefully that will permanently address these kinds of bugs. >>> > > > >>> > > > Hopefully within a few weeks there'll be a developer preview (0.89 >>> > > > series) that has that code. >>> > > > >>> > > > -ryan >>> > > > >>> > > > On Tue, Jul 20, 2010 at 2:52 PM, Karthik Kambatla >>> > > > <[email protected]> wrote: >>> > > > > Hi all >>> > > > > >>> > > > > I have been noticing a slightly flaky behavior with respect to >>> > > > > HBaseAdmin.tableExists(). After I have created the table, it >>> > > returns >>> > > > > true/false every time when called. In hbase shell, the list >>> > command >>> > > > prints >>> > > > > out all the tables sometimes, and shows no tables on other >>> > > occasions. Any >>> > > > > ideas as to why this happens? >>> > > > > >>> > > > > The master on my cluster runs the HDFS - namenode, secondary >>> > > namenode; >>> > > > HBase >>> > > > > - HMaster, HQuorumPeer. The 12 slaves run the HDFS datanode and >>> > > HBase >>> > > > > HRegionServer. Below is the transcript of operations on hbase >>> > > shell... >>> > > > Every >>> > > > > time I exit and enter hbase shell, the output changes. >>> > > > > >>> > > > > Thanks >>> > > > > Karthik >>> > > > > >>> > > > > hbase(main):002:0> count 'SUBSCRIPTIONS' >>> > > > > 0 row(s) in 6.0560 seconds >>> > > > > hbase(main):003:0> create 'Test' >>> > > > > 0 row(s) in 1.0840 seconds >>> > > > > hbase(main):004:0> list >>> > > > > SUBSCRIPTIONS >>> > > > > >>> > > > > >>> > > > > 1 row(s) in 0.0090 seconds >>> > > > > >>> > > > > hbase(main):010:0> create 'Test' >>> > > > > NativeException: org.apache.hadoop.hbase.TableExistsException: >>> > > > > org.apache.hadoop.hbase.TableExistsException: Test >>> > > > > at >>> > > > >>> > org.apache.hadoop.hbase.master.HMaster.createTable(HMaster.java:798) >>> > > > > at >>> > > > >>> > org.apache.hadoop.hbase.master.HMaster.createTable(HMaster.java:762) >>> > > > > at sun.reflect.GeneratedMethodAccessor13.invoke(Unknown >>> > Source) >>> > > > > at >>> > > > > >>> > > > >>> > > >>> > sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccesso >>> > > rImpl.java:25) >>> > > > > at java.lang.reflect.Method.invoke(Method.java:597) >>> > > > > at >>> > > org.apache.hadoop.hbase.ipc.HBaseRPC$Server.call(HBaseRPC.java:657) >>> > > > > at >>> > > > > >>> > > >>> > org.apache.hadoop.hbase.ipc.HBaseServer$Handler.run(HBaseServer.java:91 >>> > > 5) >>> > > > > >>> > > > > >>> > > > > hbase(main):011:0> list >>> > > > > SUBSCRIPTIONS >>> > > > > >>> > > > > 1 row(s) in 0.0070 seconds >>> > > > > hbase(main):012:0> exit >>> > > > > >>> > > > > [kamba...@mercado-2 ~]$ hbase shell >>> > > > > HBase Shell; enter 'help<RETURN>' for list of supported commands. >>> > > > > Version: 0.20.5, rUnknown, Thu Jul 15 22:27:05 PDT 2010 >>> > > > > hbase(main):001:0> list >>> > > > > TOPICS >>> > > > > >>> > > > > Test >>> > > > > >>> > > > > TestTable >>> > > > > >>> > > > > USERS >>> > > > > >>> > > > > 4 row(s) in 0.0930 seconds >>> > > > > hbase(main):005:0> exit >>> > > > > >>> > > > > [kamba...@mercado-2 ~]$ hbase shell >>> > > > > HBase Shell; enter 'help<RETURN>' for list of supported commands. >>> > > > > Version: 0.20.5, rUnknown, Thu Jul 15 22:27:05 PDT 2010 >>> > > > > hbase(main):001:0> list >>> > > > > EVENTS >>> > > > > >>> > > > > SUBSCRIPTIONS >>> > > > > >>> > > > > TOPICS >>> > > > > >>> > > > > TestTable >>> > > > > >>> > > > > USERS >>> > > > > >>> > > > > 5 row(s) in 0.1000 seconds >>> > > > > >>> > > > >>> >> >> >
