Enterprise Replication backup question...
A DBA asked whether it's acceptable that his IDS 9.4/AIX Enterprise Replication target (secondary) servers get no ontape/onbar backups at all, with recovery relying on re-initialising the instance (oninit -i) and reloading tables unloaded from the primary. Replies agreed this is workable but poor practice (it was apparently done originally to save disk space): with ER, unlike HDR, each server is an independent instance and should be backed up. Recommendation was to archive all ER participants, staggering full backups across servers (e.g. primary Sunday, secondary A Tuesday, secondary B Thursday) so a more recent archive from another server can speed recovery. The poster accepted this and planned to change the setup.
Auto-generated by DrWatson from the posts below — may be imperfect; read the full thread.
Hi everyone. I have some IDS 9.4 FC2X1 instances running in AIX 5.2 that have
Enterprise Replication installed. On the primary servers we do typical full
ontape backups weekly with incrementals daily but on the secondary servers
(the replicate servers) no backups are being performed-none at all. This
system has been running like this for a number of years and to date there has
never been needed a full restore (just lucky maybe). So we are in position to
completely recover the primary servers but all we have to recover the
secondary servers would be to unload the tables from the primary, transfer
them to the secondary servers and reload them. Now this works fine for the
application tables. But if we had a hardware/disk crash we would have no other
way to recover the secondary server sysmaster database. So....that means,
unless I am missing something...that the recovery "plan" that was set up on
these servers many years ago in the event of a hardware/disk crash would be to
A) recover the primary servers from full backups (that's fine), B) unload the
replicate tables to the secondary servers and then C) RECREATE THE SECONDARY
instances by completing recreating them with ONINIT -i and then after the
newly created secondary instances are recreated D) reload the tables from the
primary servers.
My question here is...is this a typical setup that others have seen? Is it
"safe"? Is it ok not to be performing actual ONTAPE (or ONBAR) backups on the
secondary servers? I am a little concerned that should we have a serious
disaster situation that all I would have as a recovery "plan" on the secondary
servers would be to re-init (oninit -i) the instances and that may not be the
best way.
Thanks, Will
This is to answer my own question. I believe the original dba that set this up simply decided not to do backups on the secondary server due to space limitations. Hence this is not a good design and will need to be addressed. Thanks. Will
In case of Enterprise Replication, I would recommend you take full backup
of your secondary (second) instance. It's lot easier to recover from back
up than performing the tasks you listed.
On the other hand, if were using High Availability Data-Replication then
you don't need to take backup of secondary as secondary server itself is a
full replica of primary server.
Regards,
Nilesh
ids-bounces@iiug.org wrote on 04/09/2009 10:17:13 AM:
> [image removed]
>
> Enterprise Replication backup question... [15507]
>
> WILL LANDSTROM
>
> to:
>
> ids
>
> 04/09/2009 10:18 AM
>
> Sent by:
>
> ids-bounces@iiug.org
>
> Please respond to ids
>
> Hi everyone. I have some IDS 9.4 FC2X1 instances running in AIX 5.2 that
have
> Enterprise Replication installed. On the primary servers we do typical
full
> ontape backups weekly with incrementals daily but on the secondary
servers
> (the replicate servers) no backups are being performed-none at all. This
> system has been running like this for a number of years and to date there
has
> never been needed a full restore (just lucky maybe). So we are in
position to
> completely recover the primary servers but all we have to recover the
> secondary servers would be to unload the tables from the primary,
transfer
> them to the secondary servers and reload them. Now this works fine for
the
> application tables. But if we had a hardware/disk crash we would
> have no other
> way to recover the secondary server sysmaster database. So....that means,
> unless I am missing something...that the recovery "plan" that was set up
on
> these servers many years ago in the event of a hardware/disk crash
> would be to
> A) recover the primary servers from full backups (that's fine), B) unload
the
> replicate tables to the secondary servers and then C) RECREATE THE
SECONDARY
> instances by completing recreating them with ONINIT -i and then after the
> newly created secondary instances are recreated D) reload the tablesfrom
the
> primary servers.
>
> My question here is...is this a typical setup that others have seen? Is
it
> "safe"? Is it ok not to be performing actual ONTAPE (or ONBAR) backups on
the
> secondary servers? I am a little concerned that should we have a serious
> disaster situation that all I would have as a recovery "plan" on
thesecondary
> servers would be to re-init (oninit -i) the instances and that may not be
the
> best way.
>
> Thanks, Will
>
>
>
*******************************************************************************
> Forum Note: Use "Reply" to post a response in the discussion forum.
>
In the purest sense the plan is "safe", it is just not sane. Recovery of
one of the secondary servers (or even of the primary if the latest full
backup were damaged) will take longer than necessary.
You should be backing up all of your Enterprise Replication servers on a
similar schedule (so maybe you perform a full backup of the primary server
on Sunday, of secondary A on Tuesday, and of secondary B on Thursday) so
that in a pinch you can restore a more recent archive from another server to
save the recovery time of rolling forward logical logs.
Come attend my Informix Best Practices session at the IIUG Conference in
Overland Park, KS on April 27.
Art
Art S. Kagel
Oninit (www.oninit.com)
IIUG Board of Directors (art@iiug.org)
Disclaimer: Please keep in mind that my own opinions are my own opinions and
do not reflect on my employer, Oninit, the IIUG, nor any other organization
with which I am associated either explicitly or implicitly. Neither do
those opinions reflect those of other individuals affiliated with any entity
with which I am affiliated nor those of the entities themselves.
On Thu, Apr 9, 2009 at 11:17 AM, WILL LANDSTROM <willlandstrom@yahoo.com>wrote:
> Hi everyone. I have some IDS 9.4 FC2X1 instances running in AIX 5.2 that
> have
> Enterprise Replication installed. On the primary servers we do typical full
> ontape backups weekly with incrementals daily but on the secondary servers
> (the replicate servers) no backups are being performed-none at all. This
> system has been running like this for a number of years and to date there
> has
> never been needed a full restore (just lucky maybe). So we are in position
> to
> completely recover the primary servers but all we have to recover the
> secondary servers would be to unload the tables from the primary, transfer
> them to the secondary servers and reload them. Now this works fine for the
> application tables. But if we had a hardware/disk crash we would have no
> other
> way to recover the secondary server sysmaster database. So....that means,
> unless I am missing something...that the recovery "plan" that was set up on
> these servers many years ago in the event of a hardware/disk crash would be
> to
> A) recover the primary servers from full backups (that's fine), B) unload
> the
> replicate tables to the secondary servers and then C) RECREATE THE
> SECONDARY
> instances by completing recreating them with ONINIT -i and then after the
> newly created secondary instances are recreated D) reload the tables from
> the
> primary servers.
>
> My question here is...is this a typical setup that others have seen? Is it
> "safe"? Is it ok not to be performing actual ONTAPE (or ONBAR) backups on
> the
> secondary servers? I am a little concerned that should we have a serious
> disaster situation that all I would have as a recovery "plan" on the
> secondary
> servers would be to re-init (oninit -i) the instances and that may not be
> the
> best way.
>
> Thanks, Will
>
>
>
>
*******************************************************************************
> Forum Note: Use "Reply" to post a response in the discussion forum.
>
>
--001636af034291374a046751d805
Art. Thanks for the response. I do actually plan on attending some of your seminars in 2009. Will
Great! I'll see you in Overland Park then. Art Art S. Kagel Oninit (www.oninit.com) IIUG Board of Directors (art@iiug.org) Disclaimer: Please keep in mind that my own opinions are my own opinions and do not reflect on my employer, Oninit, the IIUG, nor any other organization with which I am associated either explicitly or implicitly. Neither do those opinions reflect those of other individuals affiliated with any entity with which I am affiliated nor those of the entities themselves. On Wed, Apr 15, 2009 at 9:25 AM, WILL LANDSTROM <willlandstrom@yahoo.com>wrote: > Art. Thanks for the response. I do actually plan on attending some of your > seminars in 2009. > Will > > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > --0016e64135de38590c0467c0d3c7