I've put together a quick gist to reproduce the problem:
Here are some more notes:
I can reliably reproduce the problem if I create the index, insert the 2
records in the gist, then stop elasticsearch, restart it, and try to
execute the upsert statement.
When the VerifyError occurs, the CURL request just hangs.
Interestingly, it appears to happen only on a document with a large
number of nested docs, while it doesn't appear to happen with a doc
containing a few nested docs
Even for the upsert that causes the VerifyError the first time, if I
cancel the CURL request then run it again, it works properly.
The main reason this is an issue is that we are running similar upsert
statements as this within and elasticsearch river. When the VerifyError
occurs, the whole thread seems to block indefinitely and the river stops
processing anything.
It's possible as suggested in other posts, that it's a bug in MVEL, but it
seems that i'm not using MVEL for my scripts (although possibly the script
I'm using gets compiled to MVEL at a later time?).
Here are some other details:
I first noticed this issue on a production AWS machine running:
latest master version of Elasticsearch from github (I'm using the latest
from github in case it's been fixed or is using a more recent version of
MVEL which fixes the problem)
Any assistance would be appreciated. I'll continue to debug and see if I
can get any more info on the cause or a fix.
thanks.
I'm having an intermittent problem with an upsert statement. The issue
appears to be in compiling the script, but only occurs under particular
circumstances.
I've seen two similar issues reported, but I believe this is slightly
different as it's using the standard script, not an MVEL script.
These reports mention a very similar stacktrace to what I'm seeing:
I can reliably reproduce the problem if I create the index, insert the 2
records in the gist, then stop elasticsearch, restart it, and try to
execute the upsert statement.
When the VerifyError occurs, the CURL request just hangs.
Interestingly, it appears to happen only on a document with a large
number of nested docs, while it doesn't appear to happen with a doc
containing a few nested docs
Even for the upsert that causes the VerifyError the first time, if I
cancel the CURL request then run it again, it works properly.
The main reason this is an issue is that we are running similar upsert
statements as this within and elasticsearch river. When the VerifyError
occurs, the whole thread seems to block indefinitely and the river stops
processing anything.
It's possible as suggested in other posts, that it's a bug in MVEL, but it
seems that i'm not using MVEL for my scripts (although possibly the script
I'm using gets compiled to MVEL at a later time?).
Here are some other details:
I first noticed this issue on a production AWS machine running:
latest master version of Elasticsearch from github (I'm using the latest
from github in case it's been fixed or is using a more recent version of
MVEL which fixes the problem)
Any assistance would be appreciated. I'll continue to debug and see if I
can get any more info on the cause or a fix.
thanks.
no probs, I'll update if I find out any more.
thanks for the quick response!
regards
Bob.
On Friday, May 10, 2013 9:43:00 PM UTC+10, Alexander Reelsen wrote:
Hey
thanks a lot for taking your time to create a reproducable report. I've
created an issue at https://github.com/elasticsearch/elasticsearch/issues/3021
Please add anything important there, if you find out more or I described
something incorrectly. Thanks a lot!
--Alex
On Fri, May 10, 2013 at 3:53 AM, rockbobsta <b...@figjamit.com.au<javascript:>
wrote:
Hi,
I'm having an intermittent problem with an upsert statement. The issue
appears to be in compiling the script, but only occurs under particular
circumstances.
I've seen two similar issues reported, but I believe this is slightly
different as it's using the standard script, not an MVEL script.
These reports mention a very similar stacktrace to what I'm seeing:
I can reliably reproduce the problem if I create the index, insert the
2 records in the gist, then stop elasticsearch, restart it, and try to
execute the upsert statement.
When the VerifyError occurs, the CURL request just hangs.
Interestingly, it appears to happen only on a document with a large
number of nested docs, while it doesn't appear to happen with a doc
containing a few nested docs
Even for the upsert that causes the VerifyError the first time, if I
cancel the CURL request then run it again, it works properly.
The main reason this is an issue is that we are running similar upsert
statements as this within and elasticsearch river. When the VerifyError
occurs, the whole thread seems to block indefinitely and the river stops
processing anything.
It's possible as suggested in other posts, that it's a bug in MVEL, but
it seems that i'm not using MVEL for my scripts (although possibly the
script I'm using gets compiled to MVEL at a later time?).
Here are some other details:
I first noticed this issue on a production AWS machine running:
latest master version of Elasticsearch from github (I'm using the
latest from github in case it's been fixed or is using a more recent
version of MVEL which fixes the problem)
Any assistance would be appreciated. I'll continue to debug and see if I
can get any more info on the cause or a fix.
thanks.
--
You received this message because you are subscribed to the Google Groups
"elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an
email to elasticsearc...@googlegroups.com <javascript:>.
For more options, visit https://groups.google.com/groups/opt_out.
no probs, I'll update if I find out any more.
thanks for the quick response!
regards
Bob.
On Friday, May 10, 2013 9:43:00 PM UTC+10, Alexander Reelsen wrote:
Hey
thanks a lot for taking your time to create a reproducable report. I've
created an issue at https://github.com/elasticsearch/elasticsearch/
issues/3021 https://github.com/elasticsearch/elasticsearch/issues/3021
Please add anything important there, if you find out more or I described
something incorrectly. Thanks a lot!
I can reliably reproduce the problem if I create the index, insert the
2 records in the gist, then stop elasticsearch, restart it, and try to
execute the upsert statement.
When the VerifyError occurs, the CURL request just hangs.
Interestingly, it appears to happen only on a document with a large
number of nested docs, while it doesn't appear to happen with a doc
containing a few nested docs
Even for the upsert that causes the VerifyError the first time, if I
cancel the CURL request then run it again, it works properly.
The main reason this is an issue is that we are running similar upsert
statements as this within and elasticsearch river. When the VerifyError
occurs, the whole thread seems to block indefinitely and the river stops
processing anything.
It's possible as suggested in other posts, that it's a bug in MVEL, but
it seems that i'm not using MVEL for my scripts (although possibly the
script I'm using gets compiled to MVEL at a later time?).
Here are some other details:
I first noticed this issue on a production AWS machine running:
latest master version of Elasticsearch from github (I'm using the
latest from github in case it's been fixed or is using a more recent
version of MVEL which fixes the problem)
Any assistance would be appreciated. I'll continue to debug and see if I
can get any more info on the cause or a fix.
thanks.
--
You received this message because you are subscribed to the Google
Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send
an email to elasticsearc...@**googlegroups.com.
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.