EOF exception seems like a generic exception for me. I should find the
underlying problem within my infrastructure.
26 Ağustos 2013 Pazartesi tarihinde Walter Underwood
adlı kullanıcı şöyle yazdı:
> We use Amazon EC2 machines with 34GB of memory (m2.2xlarge). The Solr
heap is 8GB. We have several
We use Amazon EC2 machines with 34GB of memory (m2.2xlarge). The Solr heap is
8GB. We have several cores, totaling about 14GB on disk. This configuration
allows 100% of the indexes to be in file buffers.
wunder
On Aug 26, 2013, at 9:57 AM, Furkan KAMACI wrote:
> Hi Walter;
>
> You said you ar
Hi Walter;
You said you are caching your documents. What is average Physical Memory
usage of your Solr Nodes?
2013/8/26 Walter Underwood
> It looks lik that error happens when reading XML from an HTTP request. The
> XML ends too soon. This should be unrelated to file buffers.
>
> wunder
>
> On
It looks lik that error happens when reading XML from an HTTP request. The XML
ends too soon. This should be unrelated to file buffers.
wunder
On Aug 26, 2013, at 9:17 AM, Furkan KAMACI wrote:
> It has a 48 GB of RAM and index size is nearly 100 GB at each node. I have
> CentOS 6.4. While index
It has a 48 GB of RAM and index size is nearly 100 GB at each node. I have
CentOS 6.4. While indexing I got that error and I am suspicious about that
it is because of high percentage of Physical Memory usage.
ERROR - 2013-08-21 22:01:30.979; org.apache.solr.common.SolrException;
java.lang.RuntimeE
What is the precise error? What kind of machine?
File buffers are a robust part of the OS. Unix has had file buffer caching for
decades.
wunder
On Aug 26, 2013, at 1:37 AM, Furkan KAMACI wrote:
> Hi Walter;
>
> You are right about performance. However when I index documents on a
> machine tha
Hi Walter;
You are right about performance. However when I index documents on a
machine that has a high percentage of Physical Memory usage I get EOF
errors?
2013/8/26 Walter Underwood
> On Aug 25, 2013, at 1:41 PM, Furkan KAMACI wrote:
>
> > Sometimes Physical Memory usage of Solr is over %9
On Aug 25, 2013, at 1:41 PM, Furkan KAMACI wrote:
> Sometimes Physical Memory usage of Solr is over %99 and this may cause
> problems. Do you run such kind of a command periodically:
>
> sudo sh -c "sync; echo 3 > /proc/sys/vm/drop_caches"
>
> to force dropping caches of machine that Solr runs a
One of my Solr Nodes at SolrCloud (4.2.1) was down for a long time. I
restarted Solr and after recovery its Physical Memory usage is 99.5% and
does not decrease. Thats why I asked that question (I don't know is it
usual that Physical Memory did not decreased for 3 days. Why my CentOS 6.4
does not
