Would I be correct in my reading of the chats on the subject, that disk-based storage for
Cluster will make Cluster the preferred solution to multi-master "replication" needs?
-J
-----Original Message-----
From: Mikael Ronström [mailto:mikael@stripped]
Sent: Monday, November 14, 2005 4:03 PM
To: Mark Kirkwood
Cc: Simon Garner; pekka@stripped; Jason Williams; cluster@stripped; Adam Dixon
Subject: Re: Disk-based storage...5.1?
Hi,
2005-11-14 kl. 05.01 skrev Mark Kirkwood:
> pekka@stripped wrote:
>>> Which is totally unclear as to exactly how it will work. eg is it
>>> whole tables only on disk, partitions of tables etc.
>> Given _attributes_ on disk. No indexed attributes in 5.1.
>> In particular PK must be in memory.
>
> Is this the status of 5.1 at this time (i.e disk based storage is
> functional)?
>
> The reason being, I checked out 5.1(.3) to examine the new clustering
> and partitioning features, However loading a table bigger than my
> allocated datamemory is not working for me (see below). If this should
> work, or is worth investigating further, then let me know and I'll log
> a bug with the details
>
No, the disk data part is not functional in the 5.1 tree yet. There is still some few
things to do. When and if those goes into the 5.1 is as usual a matter of future
decisions. Variable sized storage is however working in the current 5.1 clone which means
that VARCHAR fields now only store the actual size + 2 length bytes. When it arrives
tables will be either fully main memory or disk based. If they are disk-based, then all
fields that are not part of any index will be stored in the disk part and all fields part
of indexes will be stored in the main memory part. We might add the ability to specify
that certain non-indexed fields also goes into main memory but this can be easily
accomplished by simply adding them onto an index.
Rgrds Mikael
> Cheers
>
> Mark
>
> --
>
> mysql > LOAD DATA LOCAL INFILE '/data0/dump/fact0.dat'
> > INTO TABLE fact0 FIELDS TERMINATED BY ',';
>
> ERROR 1297 (HY000): Got temporary error 4010 'Node failure caused
> abort of transaction' from NDBCLUSTER
>
> $ ndb_3_error.log
>
> Current byte-offset of file-pointer is: 568
>
>
> Time: Monday 14 November 2005 - 11:27:01
> Status: Temporary error, restart node
> Message: Internal program error (failed ndbrequire) (Internal error,
> programming error or missing error message, please report a bug)
> Error: 2341
> Error data: DbtupIndex.cpp
> Error object: DBTUP (Line: 67) 0x0000000a
> Program: /usr/local/mysql/5.1/libexec/ndbd
> Pid: 16765
> Trace: /data0/myndb/5.1/ndb_3_trace.log.1
> Version: Version 5.1.3 (alpha)
> ***EOM***
>
>
> --
> MySQL Cluster Mailing List
> For list archives: http://lists.mysql.com/cluster
> To unsubscribe:
> http://lists.mysql.com/cluster?unsub=1
>
Mikael Ronstrom, Senior Software Architect
MySQL AB, www.mysql.com
Jumpstart your cluster:
http://www.mysql.com/consulting/packaged/cluster.html
My blog:
http://mikaelronstrom.blogspot.com
!SIG:4379178a84751861067886!
