hi,
we have the following issue on our 3.10 development cluster.
we are doing regular repairs with thelastpickle's fork of creaper.
sometimes the repair (it is a full repair in that case) hangs because
of a stuck validation compaction
nodetool compactionstats gives me
You should connect to the node with JConsole and see where the compaction
thread is stuck
2017-04-13 8:34 GMT+02:00 Roland Otta :
> hi,
>
> we have the following issue on our 3.10 development cluster.
>
> we are doing regular repairs with thelastpickle's fork of
i did 2 restarts before which did not help
after that i have set for testing purposes file_cache_size_in_mb: 0 and
buffer_pool_use_heap_if_exhausted: false and restarted again
after that it worked ... but it also could be that it just worked by accident
after the last restart and is not
What if you run it again with cache enabled?
Am 13.04.2017 12:04 schrieb "Roland Otta" :
> i did 2 restarts before which did not help
>
> after that i have set for testing purposes file_cache_size_in_mb: 0 and
> buffer_pool_use_heap_if_exhausted: false and restarted
From my original email:
http://www.mail-archive.com/user@cassandra.apache.org/msg51736.html:
> My configuration changes:
>
> listen_address:
> listen_interface_prefer_ipv6: true
listen_interface is commented out. I've just now tried again with "#
listen_interface_prefer_ipv6: false" (option
i figured out that there is an mbean
org.apache.cassandra.db.type=HintedHandoffManager with the operation
scheduleHintDelivery
i guess thats what i would need in that case. at least the docs let me
think so http://javadox.com/org.apache.cassandra/cassandra-all/3.0.0/or
What I can tell you from that trace - given that this is the correct thread
and it really hangs there:
The validation is stuck when reading from an SSTable.
Unfortunately I am no caffeine expert. It looks like the read is cached and
after the read caffeine tries to drain the cache and this is
There is a nodetool command to resume hints. Maybe that helps?
Am 13.04.2017 09:42 schrieb "Roland Otta" :
> oh ... the operation is deprecated according to the docs ...
>
>
> On Thu, 2017-04-13 at 07:40 +, Roland Otta wrote:
> > i figured out that there is an mbean
I encountered this situation also once or twice but didn't succeed. I just
deleted the old hints and ran a repair
Am 13.04.2017 10:35 schrieb "Roland Otta" :
> unfortunately it does not.
>
> i guess this is intended for resuming hinted handoff handling in case it
>
Not sure what is going on there either. Roland - can you open an issue with
the information above:
https://issues.apache.org/jira/browse/CASSANDRA
On Thu, Apr 13, 2017 at 7:49 PM, benjamin roth wrote:
> What I can tell you from that trace - given that this is the correct
>
i had a closer look at the validation executor thread (i hope thats what you
meant)
it seems the thread is always repeating stuff in
org.apache.cassandra.cache.ChunkCache$CachingRebufferer.rebuffer(ChunkCache.java:235)
here is the full stack trace ...
i am sorry .. but i have no clue whats
unfortunately it does not.
i guess this is intended for resuming hinted handoff handling in case it hase
been paused with the pausehandoff before.
i have tested it (resuming .. pausing & resuming) but it has no effect on those
old hints
On Thu, 2017-04-13 at 10:27 +0200, benjamin roth wrote:
thank you guys ... i will
i just wanted to make sure that i am not doing something completely wrong
before opening an issue
br,
roland
On Thu, 2017-04-13 at 21:35 +1200, Nate McCall wrote:
Not sure what is going on there either. Roland - can you open an issue with the
information above:
oh yes .. you are absolutely right
thank you!
i will provide all the necessary info in the created jira issue
On Thu, 2017-04-13 at 15:09 +0200, benjamin roth wrote:
you should be able to find that out by scrubbing the corresponding table(s) and
see wich one hangs?
i guess the debuglog tells
Hi Martjin,
Looking at your original message:
http://www.mail-archive.com/user@cassandra.apache.org/msg51736.html
I see you edited etc/cassandra/cassandra-env.sh, by changing:
+#JVM_OPTS="$JVM_OPTS -Djava.net.preferIPv4Stack=true"
+JVM_OPTS="$JVM_OPTS -Djava.net.preferIPv6Addresses=true"
oh ... the operation is deprecated according to the docs ...
On Thu, 2017-04-13 at 07:40 +, Roland Otta wrote:
> i figured out that there is an mbean
> org.apache.cassandra.db.type=HintedHandoffManager with the operation
> scheduleHintDelivery
>
> i guess thats what i would need in that
If you restart the server the same validation completes successfully?
If not, have you tries scrubbing the affected sstables?
2017-04-13 11:43 GMT+02:00 Roland Otta :
> thank you guys ... i will
>
> i just wanted to make sure that i am not doing something completely
i made a copy and also have the permission to upload sstables for that
particular column_family
is it possible to track down which sstable of that cf is affected or should i
upload all of them?
br,
roland
On Thu, 2017-04-13 at 13:57 +0200, benjamin roth wrote:
I think thats a good
Hi,
I'm trying to do a stress test on a a table with a collection column, but
cannot figure out how to do that.
I tried
table_definition: |
CREATE TABLE list (
customer_id bigint,
items list,
PRIMARY KEY (customer_id));
columnspec:
- name: customer_id
size: fixed(64)
you should be able to find that out by scrubbing the corresponding table(s)
and see wich one hangs?
i guess the debuglog tells you which sstable is being scrubbed.
2017-04-13 15:07 GMT+02:00 Roland Otta :
> i made a copy and also have the permission to upload sstables
sorry .. i have to correct myself .. the problem still persists.
tried nodetool scrub now for the table ... but scrub is also stuck at the same
percentage
id compaction type keyspace table
completed total unit progress
Hi All,
We are in process of migrating from STCS to LCS and was just doing few reads on
line . Below is the excerpt from Datastax recommendation on data size :
Doc link :
https://docs.datastax.com/en/landing_page/doc/landing_page/planning/planningHardware.html
I found out that if the WTEs occur, there was already another process
inserting the same primary key because I found duplicates in some places
that perfectly match the WTE logs.
Does anybody know, why this throws a WTE instead of returning [applied]' =
false ?
This is quite confusing!
2017-04-12
I think thats a good reproduction case for the issue - you should copy the
sstable away for further testing. Are you allowed to upload the broken
sstable to JIRA?
2017-04-13 13:15 GMT+02:00 Roland Otta :
> sorry .. i have to correct myself .. the problem still persists.
You are welcome Martjin! Glad to have been able to help (
Best,
-Razi
On 4/13/17, 12:13 PM, "Martijn Pieters" wrote:
On 13/04/2017, 15:06, "Khaja, Raziuddin (NIH/NLM/NCBI) [C]"
wrote:
> Looking at your original message:
On 13/04/2017, 15:06, "Khaja, Raziuddin (NIH/NLM/NCBI) [C]"
wrote:
> Looking at your original message:
>
26 matches
Mail list logo