[
https://issues.apache.org/jira/browse/HBASE-13260?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14518355#comment-14518355
]
Enis Soztutar commented on HBASE-13260:
---------------------------------------
We need one more change to the test code. RegionStore.getNumThreads() always
returns 1, it does not track the actual number of threads that it is started
with. Thus, the region store tests are run with 1 Worker thread in the above
case.
With the fixes, I was able to get back to the previously reported numbers:
{code}
java.lang.AssertionError: Wrote 1000000 procedures with 5 threads with
useProcV2Wal=false hsync=false in 1mins, 14.735sec (74.735sec)
{code}
Not an order of magnitude, but this is still slow compared to the WAL based
one. A couple of things are of note:
- With increased concurrency, the region performs worse (>10 threads)
- Even with SKIP_WAL, the numbers does not look better. This is probably due
to the region internals and memstore append being the bottleneck rather than
the FSHLog. I did not yet spend time to understand why we are bottlenecked
around 10K puts and deletes (20K ops per sec) per sec with SKIP_WAL.
If we cannot improve the region store perf, I would be also in favor of using
the WAL based one.
> Bootstrap Tables for fun and profit
> ------------------------------------
>
> Key: HBASE-13260
> URL: https://issues.apache.org/jira/browse/HBASE-13260
> Project: HBase
> Issue Type: Bug
> Reporter: Enis Soztutar
> Assignee: Enis Soztutar
> Fix For: 2.0.0, 1.1.0
>
> Attachments: hbase-13260_bench.patch, hbase-13260_prototype.patch
>
>
> Over at the ProcV2 discussions(HBASE-12439) and elsewhere I was mentioning an
> idea where we may want to use regular old regions to store/persist some data
> needed for HBase master to operate.
> We regularly use system tables for storing system data. acl, meta, namespace,
> quota are some examples. We also store the table state in meta now. Some data
> is persisted in zk only (replication peers and replication state, etc). We
> are moving away from zk as a permanent storage. As any self-respecting
> database does, we should store almost all of our data in HBase itself.
> However, we have an "availability" dependency between different kinds of
> data. For example all system tables need meta to be assigned first. All
> master operations need ns table to be assigned, etc.
> For at least two types of data, (1) procedure v2 states, (2) RS groups in
> HBASE-6721 we cannot depend on meta being assigned since "assignment" itself
> will depend on accessing this data. The solution in (1) is to implement a
> custom WAL format, and custom recover lease and WAL recovery. The solution in
> (2) is to have the table to store this data, but also cache it in zk for
> bootrapping initial assignments.
> For solving both of the above (and possible future use cases if any), I
> propose we add a "boostrap table" concept, which is:
> - A set of predefined tables hosted in a separate dir in HDFS.
> - A table is only 1 region, not splittable
> - Not assigned through regular assignment
> - Hosted only on 1 server (typically master)
> - Has a dedicated WAL.
> - A service does WAL recovery + fencing for these tables.
> This has the benefit of using a region to keep the data, but frees us to
> re-implement caching and we can use the same WAL / Memstore / Recovery
> mechanisms that are battle-tested.
>
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
