Minimum would be to have check file system as often as possible and low watermark as low as possible. This would have staging run all the time staging the disk. Of course, you didn't say what is the capacity of the disk so I assume you wish to stage those as soon as possible. Nevertheless, staging will have impact on backup performance most probably.

Hello Hrvoje,
Thanks for the response.
We are expecting 100 to 120 GB disk. Can you please let me know how to calculate the staging parameters. This disk capacity is offcourse less than total amount of backup.

What will happened if during backup disk become full ? Can staging and backup be run in parallel.

Thanks
With Regards
Hemant.

Hello,
Can the stage and destination pool be the same?

Regards
Hemant.

Let's assume you have disk of 100GB. Your backup volume is bigger - no idea how much bigger.

You must first test what is the speed of read and write onto that disk when you have staging and backup running at the same time.

This is because you must make sure that time required to stage from hight watermark to low watermark is less than time required to fill the disk from high watermark to disk full. And you must calculate on something else; staging doesn't start automatically as you reach high watermark as NW does not run checks every millisecond. This check runs according to file system check interval set in staging resource so you must calculate that too.

In start if you do not know your full backup volume, to be safe, you can set low watermark to 1% and hgh watermark to 10% and have file system check run every 30 minutes.

Hello Hrvoje,
Thanks for information.
I still have some doubts.
1) What will happened if during backup if disk become full and the file type device goes to service mode.

2) As part of staging operation, we are doing high disk operation (i.e Writing 100 GB data and the sending it back to tape i.e read operation). Is there any restriction about the disk for example: RAID configuration, defragmentation etc


Thanks
With Regards
Hemant.

1) it won't go into service mode because of that. When full it will trigger staging automatically, but you will need to wait for disk to be staged as cleanup happens only after whole staging process has finished (and not after individual ssid has been staged)
2) No. Certain RAID configuration are more suitable for write and some more for read operation, but that's pretty much all.

100GB disk is pretty small and usually people are using much more.

True ... Basically the whole idea of staging is to have a copy on disk for faster recovery. With the tape drive technology today, backing up to disk is not much faster. Unless you have at least one copy maintained on disk with the older copies being staged out, the whole concept of staging does not make sense to me.

Well ... as for point 1), if the high water mark is reached (should be less than 100%), the stage out to tape operation would start thus clearing space.

As for 2), as it is a high i/o operation, it would impact your disk access operations if there are other applications that use the same server. There are no restrictions in terms of RAID level, etc for setting up a disk device, though there are best practices

Ok ... That's one way of working around your problem. If you have SAN in your environment, sharing the tape drives over your SAN might be a viable alternative. Just a suggestion though.

Hello Hari,
Following is the our motivation -
In the past tape-drives and tapes got damaged after a short period of operation because the Read-and-Write-processes of the tape-drives were often interrupted instead of working continuous. This is caused by a fractional data-transmission of backup data. The staging will solve the problem by buffering the fractional data transmission between Networker server and it¿s clients and by providing a continuous data stream towards tape-drives (Read and Write direction).

Regards
Hemant.

Hello All,
We need information about staging configuration.

Following is the information about Backup configuration -
===========================================================================
Group - No. of Clients - Start time - Full Backup Size/Date
============================================================================

Filesystem2 - 2 - 3 AM - 10 GB/Full on 3rd of every month, Rest Icnr
============================================================================

IncrementalDBbackup - 1 - 6 AM - 3 GB/ Incr Every day
============================================================================

Filesystem3 - 2 - 6: 30 AM - 25 GB/Full on First of very month, Rest Incr
============================================================================

Database Backup - 3 - 7 PM - 150 GB/Full on 3rd of every month, 30 GB /Rest Incr
============================================================================

Filesystem 1 - 10 - 10 PM -175 GB/Full on First of every month, Rest Incr
============================================================================



We have a maximum hard disk capacity of 500 GB for staging.


Can you please let us know how to configure the following staging parameter so that we will never face disk full condition and backup and staging goes smoothly.

1) High Water Mark
2) Low Water Mark
3) Max Storage Period
4) Recover Space Interval
5) File Check interval

Thanks
With Regards
Hemant.

1. High Water mark - 90
2. Low Water mark - 50
3. Max Storage - 2 days
4. Recover space interval - 12 hours
5. File check interval - 12 hours

Should work. Have taken the full backup into account and tried to come up with the figures. Would say that you need to try it out.

It's in the backup data volume and schedule for the backup. In your case, worst case, you would have 1 full of all servers at the same time. Calculate the space requirement. Then see if you can accomodate more. Would help you get the number of days. High water mark and low water mark would depend on the amount of time required to stage out a backup vs the speed at which data is coming in and how long you want to keep data in your staging area. In this case, I just made a judgement call from past experiance. As for the other two, that would depend on the frequency of your backup. No rocket science really.

Hello Hari,
Thanks for the response.
We don't want backup to remain more than 1 day. so we will set Max Storage to 1 day. but i would like to know how you reached/estimate/calculate this figure.




Regards
Hemant Joshi
Blr.

Hello All,
I was doing some test with staging.

1) I started the backup and during backup the disk become full.
There was following error messages in the Networker log.
---------------------------------------------------------------------------------------------------
Backup continures (Save set "1144494949" ) and staging (of compled save set) started. But disk marked as Full and unmounted automatically.
Eorror in the Networker UI.
12/04/08 15:26:01 nsrd: media alert event: Waiting for 1 writable volumes to backup pool.
Backup compledted in the networker log --
12/04/08 16:24:40 nsrd: media emergency: inactive ssid 1144494949
12/04/08 16:24:40 nsrd: cloning session:1 of 1 save set(s) reading from d1barsrv.d1bar.scs.com.018 6911 MB of 6911 MB
12/04/08 16:24:40 nsrd: cloning session:save sets done reading 6911 MB
12/04/08 16:24:40 nsrd: d1barsrv.d1bar.scs.com:cloning session done saving
12/04/08 16:24:40 ansrd: ansrd_clone FAILED: errnum is -7 and errstr is Bad or missing record in save set 1144494949, lost 6894787987 bytes starting at offset 0.
12/04/08 16:24:40 ansrd: failed to execute MODE_CLONE
nsrstage: error, Bad or missing record in save set 1144494949, lost 6894787987 bytes starting at offset 0.
nsrstage: Cannot open nsrstage session with d1barsrv.d1bar.scs.com
-------------------------------------------------------------------------------------------------------------------------
I noticed that the file system device is unmounted automatically. I mounted the device again the new backup goes fine and staging works fine.

1) The "1144494949" save set still in the file system device. How to delete this save set using networker.

2) What "Bad or missing record in save set " indicates. Is it mean that the backup not completed successfully and there is data lose.


Regards
Hemant Joshi.