BTW, you can use d...@hbase.apache.org rather than issues@. The latter is
more for emails from jira.
On Mon, May 2, 2016 at 7:11 PM, Enis Söztutar <e...@apache.org> wrote:
> Thanks for reporting.
>
> In master and branch-1, this part of the code is very different and no
> longer has the problem.
>
> Did you check the latest 0.98 code base? It may not be worth fixing at
> this point.
>
> Enis
>
> On Mon, May 2, 2016 at 6:21 AM, WangYQ <wangyongqiang0...@163.com> wrote:
>
>> the code :
>>
>> private synchronized CatalogTracker getCatalogTracker()
>> throws ZooKeeperConnectionException, IOException {
>> CatalogTracker ct = null;
>> try {
>> ct = new CatalogTracker(this.conf);
>> ct.start();
>> } catch (InterruptedException e) {
>> // Let it out as an IOE for now until we redo all so tolerate IEs
>> Thread.currentThread().interrupt();
>> throw new IOException("Interrupted", e);
>> }
>> return ct;
>> }
>>
>>
>> I think we can make CatalogTracker be a object of HBaseAdmin class, can
>> reduce many object create and destroy, reduce client to ZK
>>
>>
>>
>>
>>
>>
>>
>> At 2016-04-19 21:09:42, "WangYQ" <wangyongqiang0...@163.com> wrote:
>>
>> in hbase 0.98.10, class "HBaseAdmin "
>> line 303, method "tableExists", will create a catalogTracker for
>> every call
>>
>>
>> we can let a HBaseAdmin object use one CatalogTracker object, to reduce
>> the object create, connect zk and so on
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>
>
