Hi Bruno
Regarding the perf issue with more CPs you might be interested in seeing
this
https://issues.apache.org/jira/browse/HBASE-25277
Regards
Ram
On Thu, Nov 5, 2020 at 11:25 PM ramkrishna vasudevan <
ramkrishna.s.vasude...@gmail.com> wrote:
> This is great information. Thanks Bruno. Keep
This is great information. Thanks Bruno. Keep us posted if you find
anything more when you have your isolated tests done.
Regards
Ram
On Thu, Nov 5, 2020 at 10:36 PM Bruno Dumon wrote:
> Hi Ram,
>
> Sorry I missed your question. Yes, we do use addColumns.
>
> Specifically, we also query the
Hi Ram,
Sorry I missed your question. Yes, we do use addColumns.
Specifically, we also query the Phoenix existence marker, which is a column
named "_0", which is typically the last column in the row (we query this
through custom scans, not through Phoenix). It does seem that reading the
last
Hi Bruno/Jan
Just a query here. I read your emails in this thread. One simple question
just to ensure if your tests were similar to Andrew's test. Did your scan
query have addColumns explicitly added covering all (or most of) the
columns in the rows?
Regards
Ram
On Wed, Jul 15, 2020 at 7:02
This is a good finding, nice work!
I added a comment on HBASE-24742 that mentions HBASE-24637 on the off
chance they are related, although I suspect more changes are implicated by
the 2.x regression.
On Tue, Jul 14, 2020 at 5:53 PM Bharath Vissapragada
wrote:
> FYI, we filed this today
I went out on vacation (and am still out) before tracking this down. If you
are waiting for me to make more progress with HBASE-24637, I can do that in
a couple of weeks, Anyone is welcome to step in sooner. ..
On Tue, Jul 14, 2020 at 11:38 AM Josh Elser wrote:
> Wow. Great stuff, Andrew!
>
>
FYI, we filed this today https://issues.apache.org/jira/browse/HBASE-24742.
We ran into a similar regression when upgrading from 1.3 based branch to
1.6 based branch. After some profiling and code analysis we narrowed down
the code paths.
On Tue, Jul 14, 2020 at 11:38 AM Josh Elser wrote:
>
Wow. Great stuff, Andrew!
Thank you for compiling and posting it all here. I can only imagine how
time-consuming this was.
On 6/26/20 1:57 PM, Andrew Purtell wrote:
Hey Anoop, I opened https://issues.apache.org/jira/browse/HBASE-24637 and
attached the patches and script used to make the
Hey Anoop, I opened https://issues.apache.org/jira/browse/HBASE-24637 and
attached the patches and script used to make the comparison.
On Fri, Jun 26, 2020 at 2:33 AM Anoop John wrote:
> Great investigation Andy. Do you know any Jiras which made changes in SQM?
> Would be great if you can
Great investigation Andy. Do you know any Jiras which made changes in SQM?
Would be great if you can attach your patch which tracks the scan flow. If
we have a Jira for this issue, can you pls attach?
Anoop
On Fri, Jun 26, 2020 at 1:56 AM Andrew Purtell
wrote:
> Related, I think I found a
Related, I think I found a bug in branch-1 where we don’t heartbeat in the
filter all case until we switch store files, so scanning a very large store
file might time out with client defaults. Remarking on this here so I don’t
forget to follow up.
> On Jun 25, 2020, at 12:27 PM, Andrew
I repeated this test with pe --filterAll and the results were revealing, at
least for this case. I also patched in thread local hash map for atomic
counters that I could update from code paths in SQM, StoreScanner,
HFileReader*, and HFileBlock. Because a RPC is processed by a single
handler thread
12 matches
Mail list logo
