From: Johan Andersson
Date: October 21 2010 6:07am
Subject: Re: All tables remove on ndbd --initial
List-Archive: http://lists.mysql.com/cluster/7854
Message-Id: <44F0C75F-D00F-434B-8484-3F7F42C7B2B4@severalnines.com>
MIME-Version: 1.0 (iPad Mail 7B367)
Content-Type: text/plain;
	charset=us-ascii
Content-Transfer-Encoding: quoted-printable

Hi,
Leave these in the my.cnf file, it is mysql server specific settings.

Br
Johan

Sent from my iPad

On 21 okt 2010, at 06.45, "Tears !" <unix.co@stripped> wrote:

> Dear Reza and Johan,
>=20
> I am running another MySQL 5.1 server without cluster. For the tunning
> purpose I have set these option in my.cnf.
>=20
> Now my question is, can I use the same setting on my both storage =
nodes
> under mysqld section?
>=20
> Here is the list of that option.
>=20
> *skip-name-resolve*
> *federated*
> *interactive_timeout =3D 120*
> *wait_timeout =3D 120*
> *key_buffer_size =3D 512M*
> *max_connections =3D 800*
> *max_connect_errors =3D 100*
> *max_allowed_packet =3D 16M*
> *table_open_cache =3D 2048*
> *sort_buffer_size =3D 8M*
> *join_buffer_size =3D 8M*
> *read_buffer_size =3D 8M*
> *read_rnd_buffer_size =3D 16M*
> *myisam_sort_buffer_size =3D 64M*
> *myisam_max_sort_file_size =3D 10G*
> *thread_cache_size =3D 512*
> *query_cache_size =3D 64M*
> *query_cache_limit =3D 4M*
> *# Try number of CPU's*2 for thread_concurrency*
> *thread_concurrency =3D 16*
> *ft_min_word_len =3D 3*
> *tmp_table_size =3D 128M*
> *expire_logs_days =3D 15*
>=20
> Regards,
>=20
> Umar
>=20
> On Thu, Oct 21, 2010 at 9:17 AM, Reza Iskandar Achmad <
> reza@stripped> wrote:
>=20
>> This is important procedure when you did a change in config.ini as =
Johan
>> said.
>>=20
>> Start node at a time to let the new started node to rebuild data and
>> copying
>> fragment from another node.
>>=20
>> To be straight forward. Procedure for config.ini change:
>> 1. edit config.ini
>> 2. restart all management nodes (I prefer to ndb_mgm -e nodeid stop) =
and
>> wait until ndb_mgmd exits
>> 3. re-run ndb_mgmd with --reload option, wait until management node =
started
>> 4. repeat procedure 2-3 if you have another management server
>> 5. stop first node from ndb_mgm console (e.g. ndb_mgm> 11 stop)
>> 6. wait until data node stop and another node to take over as master
>> 7. re-run first node without --initial
>> 8. wait until first node started and joined the cluster/node group. =
It
>> would
>> takes minutes or hours depends on you data size
>> 9. repeat procedure 5-8 for all data nodes
>>=20
>> After all, prepare your backup before doing any changes :) it's not =
fun to
>> lost the database
>>=20
>> Regards,
>> Reza
>>=20
>>=20
>> On 10/21/10 4:18 AM, "Johan Andersson" <johan@stripped> =
wrote:
>>=20
>>> .... And then restart _one_ data node at a time (and the mysqlds too =
if u
>> have
>>> changed any that affects them, like number of [mysqld] sections ) !
>>> Br
>>> Johan
>>>=20
>>> On 20 okt 2010, at 22:32, "Tears !" <unix.co@stripped> wrote:
>>>=20
>>>> Hi Reza, David and John,
>>>>=20
>>>> Thanks for your help, Unfortunatly i had run --initial command on =
both
>> nodes
>>>> (:.
>>>>=20
>>>> Now you people said, whenever I change something in config.ini.
>>>>=20
>>>> I should run ndb_mgmd --reload on management server ??
>>>>=20
>>>> Regards,
>>>>=20
>>>> Umar
>>>>=20
>>>> On Thu, Oct 21, 2010 at 12:45 AM, <reza@stripped> wrote:
>>>>=20
>>>>> Hello Umar,
>>>>>=20
>>>>> You should re-run ndb_mgmd with --reload option. And run ndbd =
without
>>>>> --initial.
>>>>>=20
>>>>> As long as I know, --initial option would tell ndbd to =
re-initialized
>> the
>>>>> storage. Probably there's no way to recover your table except from
>> backup.
>>>>>=20
>>>>> Hope that helps.
>>>>>=20
>>>>> Regards,
>>>>> Reza
>>>>>=20
>>>>> ------Original Message------
>>>>> From: Tears !
>>>>> To: MySQL Cluster Mailing List
>>>>> Subject: All tables remove on ndbd --initial
>>>>> Sent: Oct 21, 2010 02:31
>>>>>=20
>>>>> Dear All,
>>>>>=20
>>>>> For some reason I had changed config.ini on management node.
>>>>>=20
>>>>> After that I had kill (ndbd) on both storage nodes. And then again
>> start
>>>>> ndbd with ndbd --initial.
>>>>>=20
>>>>> But when I had checked tables in my database there was nothing , =
all
>> tables
>>>>> were removed
>>>>>=20
>>>>> Would you please help me how I can recover my old tables.
>>>>>=20
>>>>> And what is the correct procedure to change config.ini and load =
again.
>>>>>=20
>>>>> Regards,
>>>>>=20
>>>>> Umar
>>>>>=20
>>>>>=20
>>>>>=20
>>>>>=20
>>>>=20
>>>>=20
>>>> --
>>>> Umar Draz
>>>> Network Administrator
>>=20
>>=20
>>=20
>=20
>=20
> --=20
> Umar Draz
> Network Administrator