Re: [Info-Ingres] Linux LVM2 and creating a snapshot
- From: "Martin Bowes" <martin.bowes@xxxxxxxxxxxxx>
- Date: Thu, 12 Jun 2008 11:34:26 +0100
Hi David
The plethora of storage options has yet to be resolved into actual
hardware sitting in the machine room.
In all cases, the thought of throwing a 1TB database backup around isn't
thrilling me. On the current host using 15k drives in the host I back up
at a little better than 1G/minute. So a standard checkpoint would take
16hours40minutes! Then perform an OS dump to save that checkpoint...I
don't have a tape drive that can do that. I assume I'll need to get a
tape stacker as well.
The checkpoint uses gzip for compression but typically the cpu seems to
be in iowait trying to write the data to disk rather than trying to
compress the data. The checkpoint disks are in Raid 0+1 (stripped
mirror) configuration. I've experimented with dropping compression, but
didn't achieve a significantly better rate.
The database we expect to grow to 1TB is currently achieving poor
compression ratios of ~55%. But the new data is largely binary images
which have been compressed by the device capturing them. So I expect we
would see minimal compression being achieved - to the point that I'll
drop it from the template file once these images start coming online and
I see the compression ratio fall away.
I'm leaning towards stocking a SAN or NAS with 15k drives, but I've also
been told we could buy a new server with 4Tb of disk fairly cheaply.
I'm struggling with trying to work out which would give better
performance and be more useful when a recovery was required, so I've
leaned towards the SAN. However, my boss informs me that he could live
with two days of downtime. Further the hosts are largely idle on the
weekends so he's not unhappy with a 17hour backup.
Yours in confusion...
Marty
From: info-ingres-bounces@xxxxxxxxxxxxxxxxxxxxxxxxx
[mailto:info-ingres-bounces@xxxxxxxxxxxxxxxxxxxxxxxxx] On Behalf Of
David Richard
Sent: 11 June 2008 16:18
To: Ingres and related product discussion forum
Subject: Re: [Info-Ingres] Linux LVM2 and creating a snapshot
Hi Marty,
What class of storage are we talking about (Fibre 15K SCSI, 7.2K
SATA....)?
What is the RAID level (1, 1+0, dare I say 5...)?
How many 'spindles' make up the volume?
Are these 'spindles' exclusive or shared? If shared, how 'hot' are they?
How is the storage attached (Local, SAN, ISCSI, NAS...)?
Is there any intelligent cache between the disks and the host?
************************************************************************
*****
DISCLAIMER
The information contained in this e-mail is confidential and is intended
for the recipient only.
If you have received it in error, please notify us immediately by reply
e-mail and then delete it from your system. Please do not copy it or
use it for any other purposes, or disclose the content of the e-mail
to any other person or store or copy the information in any medium.
The views contained in this e-mail are those of the author and not
necessarily those of AAH Pharmaceuticals Ltd.
AAH Pharmaceuticals Ltd is a company incorporated in England and
Wales under company number 123458 and whose registered office
is at Sapphire Court, Walsgrave Triangle, Coventry, CV2 2TX
************************************************************************
*****
- Prev by Date: Re: [Info-Ingres] Linux LVM2 and creating a snapshot
- Next by Date: Re: [Info-Ingres] Linux LVM2 and creating a snapshot
- Previous by thread: Re: [Info-Ingres] Linux LVM2 and creating a snapshot
- Next by thread: Re: [Info-Ingres] Linux LVM2 and creating a snapshot
- Index(es):
Relevant Pages
|
