Storage Discussions Storage Discussions http://forums.hds.com/index.php Mon, 12 May 2008 02:24:13 -0400 5 shared NTFS volume http://forums.hds.com/index.php?showtopic=762
What I would like to do is this:
I have a SQL server backup dump of 600GB and transaction logs going to a SAN LUN on my SQL server.
I would like to attach my backup server to the same LUN and backup the SQL dump and transaction logs.
With this method none of the backup traffic is going across my LAN.

Anyone know if this solution is feasible? Or is anyone doing something similar I'm open to suggestions.

thanks

Chris T.




]]> Fri, 09 May 2008 08:56:56 -0400 762 MPIO, round_robin algorithm, reserve_policy and HDS disk http://forums.hds.com/index.php?showtopic=761
We are currently using this setup where we have the Hitachi AIX MPIO ODM and configured each disk for round_robin.
It seems that we also have to set the reserve_policy to no_reserve to allow the algorithm attribute to be set to round_robin.
It doesn't seem to like the PR_shared or PR_exclusive option as the reserve_policy.
They do seem to be defined as options in the ODM for disk/fcp/htcuspmpio and disk/fcp/htc9900mpio.

Has anyone been able to use PR_shared/PR_exclusive to get disk locking to work?
Without reserves or SCSI locks on the disks, it is possible that the SAN guys are able to de-assign the disk from the system or even assign to another system accidentally. Is there a way set these reserve policies on the disks?

Cheers,
Wing]]> Fri, 09 May 2008 03:39:56 -0400 761 Sync VS Async http://forums.hds.com/index.php?showtopic=449
He want to know if he want to use aync , Hitachi can quarantee data or not.
He know if he use sync it will slower but it ensure data intrigity. If anyone know or can recommend me what solution that it should be implement.

Customer concern about I/O performance and data intrigity. Please suggest me

Thank,
]]> Fri, 16 Feb 2007 06:18:31 -0500 449 vxstat stats vs raw hard disk IO/s http://forums.hds.com/index.php?showtopic=760
I am facing a problem which comes from question "How do you define IO ?" (IOPS).
We have hds 9570 with 12 HDDs (+2 spare drives) and two RAID-5 RGs (5D+1D). There are several LUNs on each RG. It is connected to sparc server with Solaris 10 and Veritas 4.1. We have several Oracle databases (OLTP) working on the server/array. When I run vxstat I get:
CODE

drozd:/root# vxstat -g drozd-sierra2-dg -i 60
                      OPERATIONS          BLOCKS           AVG TIME(ms)
TYP NAME              READ     WRITE      READ     WRITE   READ  WRITE

[...]

Wed May 07 15:07:13 2008
vol drozd-sierra2-vol    466239       548   7450629      7714    0.3    0.6

Wed May 07 15:08:13 2008
vol drozd-sierra2-vol    319741       693   5108947     10307    0.3    0.9

Wed May 07 15:09:13 2008
vol drozd-sierra2-vol    331786       404   5301400      4036    0.3    0.4

An average HDD can do about 200 IOPS. So if I count 200*12 I get 2400 IOPS. So how do we get 7770/s READs operatios (466239/60) ? Is the controller able to optimize the workload to achieve the result or do I mix different terms ?
]]> Wed, 07 May 2008 09:46:20 -0400 760 VMWare best practice http://forums.hds.com/index.php?showtopic=752
Platform - AMS100

Apart from the fixed path mutipathing setting in vmware, and the host modes, what else should I be looking at when configuring?

Number of esx servers pe lun?

Number of guests per lun?

Are there any docs/web sites that can be referenced?


]]> Fri, 18 Apr 2008 06:56:55 -0400 752 HDLM and Veritas Storage Foundation DMP - How to replace HDLM by SF for Windows? http://forums.hds.com/index.php?showtopic=759
we have decided to decommission HDLM in favour of Storage Foundation DMP.
I am about to retest the in place upgrade on a windows 2k3 server as the last try done by my team ended up by a complete rebuilt!

So the server is Win2k3, HDLM (2 HBAs) and 2 LUNs of 1TB each.
The LUNs are
dynamic volumes,
striped together
and presented as a mount point.

Now what would be the best sequence?
1 unmount volumes (disk mgt > delete volume...)
2 uninstall HDLM
3 reboot
4 confirm duplicated LUNs signatures
5 install Storage Foundation with DMP option
6 reboot
7 confirm consolidated LUNs signatures
8 rescan the disks and reactivate?

I will be trying it in the next days and will update with the results... but if you have any insight, please let me know.

Many Thanks

Alexandre Rappoport]]> Mon, 05 May 2008 23:23:08 -0400 759 HSSM user group http://forums.hds.com/index.php?showtopic=758
Does anyone know if there is a HSSM user group anywhere in the world. I could not find anything related to Storage Essentials at HP ITRC and considering that it is probably sold more than HSSM, its a bit disheartening to not be able to share info on this product.

I imagine at least a handful of people must have HSSM and would like to explore its abilities other than using its sluggish GUI and putting up with completely useless events.

I heard that version 6 is much better.

I also read there is a vulnerability with JBossandjetty with Storage Essentials. I wonder if HSSM has the same issue. Probably does...

http://www.ciac.org/ciac/bulletins/s-260.shtml

Stephen]]> Mon, 05 May 2008 03:59:10 -0400 758 Shadow Image with external storage http://forums.hds.com/index.php?showtopic=757
I was wondering if anyone is using Shadow Image with lesser tiered external storage. We have USP's with LDEV's that I would like to use Shadow Image to pair up on that I can present to other hosts. Thats the easy part. I just wonder about how the LDEV's on something like an externalised AMS system could cope with relatively heavy I/O especially if we were thinking about using SATA drives.

The USP uses 15k 146 GB FC disks. If I was to pair up 500 GB SATA drives, is that going to work if the USP LDEV's are working relatively hard? I believe there are options in the USP to cater for this so as not to completely muck up the cache but surely there must be something that pays.

So, perhaps, it works like this? The initial Shadow Image pair should be done during a relatively quite time. Once the pair has been created, then any changes to the data are really only at the block level and a lot of the I/O such as reads don't actually affect the pairs. Some of the True Copy resyncs are really outstanding compared to the size of the data. So, I would summise the same thing happens with SI. If I had the pair broken for only a fraction of the time, then the resync's should be ok?

Would anyone care to comment on this?

Stephen

]]> Mon, 05 May 2008 03:45:05 -0400 757 FirmWare/Microcode and Software levels http://forums.hds.com/index.php?showtopic=730
Thanks in advance,

Frank
]]> Mon, 03 Mar 2008 10:25:19 -0500 730 USP 1100 Comarison to EMC DMX-3 http://forums.hds.com/index.php?showtopic=743 As you know in DMX-3, primary installation and any main reconfiguration such as adding disk, running remote copy & … needs a new *bin.file* made by EMC.
Moreover the maintenance of DMX-3 are usually done through *Inline Command*s that are confidential so the customer is unable to support them in any levels.
I would like to know if USP 1100 installation/maintenance needs such a thing (bin.file).
Or whenever the maintenance of USP1100 needs to execute by command line( in low level), the related commands are available for customer or they are confidential (like EMC Inline Command)
Yeganeh]]> Fri, 28 Mar 2008 03:28:13 -0400 743