[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16721124#comment-16721124
]
Hudson commented on HBASE-17072:
Results for branch branch-1.3
[build #576 on
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15749575#comment-15749575
]
stack commented on HBASE-17072:
---
Ok to backport this [~busbey]
> CPU usage starts to climb up to 90-100%
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15729317#comment-15729317
]
Hudson commented on HBASE-17072:
FAILURE: Integrated in Jenkins build HBase-0.98-on-Hadoop-1.1 #1294 (See
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15729241#comment-15729241
]
Hudson commented on HBASE-17072:
FAILURE: Integrated in Jenkins build HBase-0.98-matrix #423 (See
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15724412#comment-15724412
]
stack commented on HBASE-17072:
---
Master should be ok [~apurtell] I think given its atomic ref (which ever
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15703902#comment-15703902
]
Hudson commented on HBASE-17072:
SUCCESS: Integrated in Jenkins build HBase-1.4 #548 (See
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15703215#comment-15703215
]
stack commented on HBASE-17072:
---
The follow-on is HBASE-17185
> CPU usage starts to climb up to 90-100%
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15703182#comment-15703182
]
Hadoop QA commented on HBASE-17072:
---
| (x) *{color:red}-1 overall{color}* |
\\
\\
|| Vote || Subsystem
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15702871#comment-15702871
]
stack commented on HBASE-17072:
---
Pushed to master branch... backporting.
> CPU usage starts to climb up to
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15702869#comment-15702869
]
stack commented on HBASE-17072:
---
Put up a branch-1 patch for precommit since it was painful to backport.
>
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15702625#comment-15702625
]
stack commented on HBASE-17072:
---
Thanks [~ram_krish] Will commit shortly.
On the TODO, let me finish up my
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15702539#comment-15702539
]
ramkrishna.s.vasudevan commented on HBASE-17072:
+1 since the change to get from the
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15702530#comment-15702530
]
stack commented on HBASE-17072:
---
Any +1s on this?
> CPU usage starts to climb up to 90-100% when using
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15698982#comment-15698982
]
stack commented on HBASE-17072:
---
+1 on this? I can fix the whitespace on commit.
> CPU usage starts to
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15697201#comment-15697201
]
Hadoop QA commented on HBASE-17072:
---
| (x) *{color:red}-1 overall{color}* |
\\
\\
|| Vote || Subsystem
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15692760#comment-15692760
]
Hadoop QA commented on HBASE-17072:
---
| (x) *{color:red}-1 overall{color}* |
\\
\\
|| Vote || Subsystem
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15692551#comment-15692551
]
Anoop Sam John commented on HBASE-17072:
Oh you mean we will be always passing the block on disk
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15692449#comment-15692449
]
stack commented on HBASE-17072:
---
Patch v5 restores caching of the next-blocks header. While not needed for
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15692284#comment-15692284
]
stack commented on HBASE-17072:
---
No. The order. Before patch it is seek to offset 0 twice, then to the
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15692255#comment-15692255
]
ramkrishna.s.vasudevan commented on HBASE-17072:
bq.I'm not sure why above changes file
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15691884#comment-15691884
]
Hadoop QA commented on HBASE-17072:
---
| (/) *{color:green}+1 overall{color}* |
\\
\\
|| Vote ||
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15691661#comment-15691661
]
stack commented on HBASE-17072:
---
Posted a patch that removs the above though it saves an odd seek. Needs
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15691636#comment-15691636
]
stack commented on HBASE-17072:
---
Test fails because of the below part of the patch which tries to go to the
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15690869#comment-15690869
]
stack commented on HBASE-17072:
---
bq. Normal read flow, we make use of readBlockDataInternal() and there we
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15688941#comment-15688941
]
Anoop Sam John commented on HBASE-17072:
I can see we make use of next block's size in
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15688806#comment-15688806
]
Yu Li commented on HBASE-17072:
---
Oh, regarding UT failure reported by HadoopQA, seems relative and worth a
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15688811#comment-15688811
]
stack commented on HBASE-17072:
---
Thanks you for the review [~carp84] ... yes it is a good catch by
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15688800#comment-15688800
]
Yu Li commented on HBASE-17072:
---
Nice catch and analysis!
I wonder the issue here is not only for G1GC but
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15688493#comment-15688493
]
stack commented on HBASE-17072:
---
Unpatched did 18 seeks. Patched did 17.
> CPU usage starts to climb up to
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15688488#comment-15688488
]
stack commented on HBASE-17072:
---
I ran 5x concurrent scans against my 10M table as HBASE-10676 did. There
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15688425#comment-15688425
]
Hadoop QA commented on HBASE-17072:
---
| (x) *{color:red}-1 overall{color}* |
\\
\\
|| Vote || Subsystem
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15688272#comment-15688272
]
stack commented on HBASE-17072:
---
Pushed another patch. It is just additional comments. Can I get a +1
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15688165#comment-15688165
]
stack commented on HBASE-17072:
---
Ok. Looks like I had my test counts wrong way around. I see that the
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15687520#comment-15687520
]
stack commented on HBASE-17072:
---
I ran tests overnight comparing before and after patch and it seems like
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15685436#comment-15685436
]
Hadoop QA commented on HBASE-17072:
---
| (x) *{color:red}-1 overall{color}* |
\\
\\
|| Vote || Subsystem
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15685319#comment-15685319
]
stack commented on HBASE-17072:
---
Pushed first cut at a patch. After study, Scans and random Gets are
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15681601#comment-15681601
]
stack commented on HBASE-17072:
---
bq. Why can't we keep the next block's header as a state in HFileScanner?
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15680297#comment-15680297
]
Anoop Sam John commented on HBASE-17072:
I was abt to do the math... Ya this seems too much..So
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15680085#comment-15680085
]
stack commented on HBASE-17072:
---
One thought is to do a more radical HBASE-10676 and cache block headers in
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15678305#comment-15678305
]
stack commented on HBASE-17072:
---
HBASE-10676's soln is to cache the header on the reader instance rather
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15675813#comment-15675813
]
stack commented on HBASE-17072:
---
I am in awe of the digging done here (even after second read). Seems like
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15675448#comment-15675448
]
Tomu Tsuruhara commented on HBASE-17072:
ah, I see. Thanks.
> CPU usage starts to climb up to
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15675394#comment-15675394
]
Eiichi Sato commented on HBASE-17072:
-
I think ThreadLocal#remove() needs to be called on every
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15673741#comment-15673741
]
Tomu Tsuruhara commented on HBASE-17072:
How about simply calling {{ThreadLocal#remove}} on
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15657155#comment-15657155
]
Eiichi Sato commented on HBASE-17072:
-
The problem is called [primary
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15657123#comment-15657123
]
Eiichi Sato commented on HBASE-17072:
-
Thanks for the suggestion. I wasn't aware of HBASE-17017.
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15656390#comment-15656390
]
stack commented on HBASE-17072:
---
Nice analysis. The header prefetch can save a seek when scans cross hfile
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15656376#comment-15656376
]
Anoop Sam John commented on HBASE-17072:
Ya when blocks are there in BC, we wont deal with this
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15656372#comment-15656372
]
Esteban Gutierrez commented on HBASE-17072:
---
[~sato_eiichi], take a look at HBASE-17017
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15656368#comment-15656368
]
Esteban Gutierrez commented on HBASE-17072:
---
As I think about this problem as described by
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15656321#comment-15656321
]
ramkrishna.s.vasudevan commented on HBASE-17072:
May be this is more helpful when the
[
https://issues.apache.org/jira/browse/HBASE-17072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15656297#comment-15656297
]
Anoop Sam John commented on HBASE-17072:
So the prefetching of the new block's header is not
52 matches
Mail list logo
