I have a production box running old ES "0.20.3" .. I need help in cleaning
data back from 2012 ... because although I moved the whole indexs to a
machine with 6 SSD drives ... queries got better .. but still I need to
make this thing faster
is there a *safe *way to clean (choose what to clean based on date) and
improve the performance?
On Monday, March 30, 2015 at 11:56:46 PM UTC+3, AALISHE wrote:
Hi all,
I have a production box running old ES "0.20.3" .. I need help in
cleaning data back from 2012 ... because although I moved the whole indexs
to a machine with 6 SSD drives ... queries got better .. but still I need
to make this thing faster
is there a *safe *way to clean (choose what to clean based on date) and
improve the performance?
This is a community based list and support is provided by volunteers as
their time permits. If you want timely responses then I'd recommend you
sign up for some support, we'd be happy to have you as a customer Subscriptions | Elastic Stack Products & Support | Elastic
Is all the data in the one index, or are they time based?
If you need ES to be faster then I'd suggest you upgrade, 0.20 is no longer
supported and there are a lot of performance improvements in later
versions.
On Monday, March 30, 2015 at 11:56:46 PM UTC+3, AALISHE wrote:
Hi all,
I have a production box running old ES "0.20.3" .. I need help in
cleaning data back from 2012 ... because although I moved the whole indexs
to a machine with 6 SSD drives ... queries got better .. but still I need
to make this thing faster
is there a *safe *way to clean (choose what to clean based on date) and
improve the performance?
1- yes I believe all the data is in one index .. I had attached the
index_status output ... and a screenshot for the head_plugin output
2- I am considering the upgrade to latest version .. but would that impose
a risk (if any)?
Can you create a new index and exclude what you dont need. we did this
recently because of some other mapping reason. take a snapshot (with newwer
version) before doing so.
I would rotate your index so all new data would go in to a new one
while the archive can wither and die
Maybe rotate it on a daily basis or a point of time that works for
your data sets you can still have basic aliases to summarize multiple days
depending on your query needs
Use a project called curator to purge old data on a routine basis
(cron)
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.