Bug #66894 demo_table_large does not work with default policy for large data (> 2000 bytes)
Submitted: 20 Sep 2012 15:08
Reporter: Michael Gagnon Email Updates:
Status: Open Impact on me:
None 
Category:MySQL Cluster: Memcached Severity:S3 (Non-critical)
Version:mysql-cluster-gpl-7.2.7-debian6.0-x86_64 OS:Any
Assigned to: CPU Architecture:Any

[20 Sep 2012 15:08] Michael Gagnon
Description:
Using the large_values_table with any policy other than caching does not work for data exceeding the size of the main table. With caching, storing and retrieving works properly. With any other role (eg: ndb-only), it will store correctly but not retrive correctly -- it returns no data. This is occurs with the provided sample configuration.

How to repeat:
- Run out of the box ndbmemcache setup
- start memcache with default role
- start telnet:

set b:abc 0 0 2200
<2200 bytes>
STORED

get b:abc
VALUE b:abc 0 0

END
[9 May 2013 14:18] Joshua Gigg
I'm also facing this issue.

With -vvv and debug=true:

47 - Running task: (conn_read)
47: going from conn_read to conn_parse_cmd
47 - Running task: (conn_parse_cmd)
<47 get prefix:myfirstkey
worker.0 ndb_get():creating workitem 0.1
worker.0 schedule():SchedulerWorker / config gen. 1
worker.0 worker_set_ext_flag(): 0.1: T
worker.0 --> do_read()
worker.0 --> setKeyForReading()
worker.0 workitem_allocate_rowbuffer_1(): 2028 [cls 11]
worker.0 schedule():0.1 placed on send queue.
cl0.conn0.send run_ndb_send_thread():Sent 0.1
cl0.conn0.poll run_ndb_poll_thread(): ** adding 0.1 to wait group **
cl0.conn0.poll run_ndb_poll_thread():Polling 0.1
cl0.conn0.poll callback_main():Success.
cl0.conn0.poll readFromHeader():14615758/2732 (1 parts of size 13950)
cl0.conn0.poll --> reschedule()
cl0.conn0.poll run_ndb_poll_thread():Rescheduling 0.1
cl0.conn0.send run_ndb_send_thread():Sent 0.1
cl0.conn0.poll run_ndb_poll_thread(): ** adding 0.1 to wait group **
cl0.conn0.poll run_ndb_poll_thread():Polling 0.1
cl0.conn0.poll callback_ext_parts_read(): 0.1
cl0.conn0.poll worker_close():0.1
cl0.conn0.poll --> ~ExternalValue()
Got notify from 47, status 0
<47 get prefix:myfirstkey
worker.0 ndb_get():Got read callback on workitem 0.1: Transaction succeeded
worker.0 ndb_get_item_info(): new hash_item
>47 sending key prefix:myfirstkey

>47 END
47: going from conn_parse_cmd to conn_mwrite
worker.0 release_and_free():Releasing workitem 0.1.
worker.0 --> release()
worker.0 release():Returned NdbInstance to freelist.
worker.0 ndb_release():Releasing a hash item.
47: going from conn_mwrite to conn_new_cmd
47: going from conn_new_cmd to conn_waiting
47: going from conn_waiting to conn_read

Also appears to be happening to someone @ http://lists.mysql.com/cluster/8433