24x7 vs cold backup

[chauvin]

We are currently desciding of the backup strategy.

I wanted to know if anyone experienced problems with the 24x7 hot backup when using it during high activity, as it is said in the doc. SC will be used for a 24/7 worldwide support, so we don't know yet if there will be any periode in the day with low activity.



Here follows the choosen achitecture :



SC4 : 2 x IBM pSeries660 (AIX, HACMP)

P4 : SAN (storage area network) with mirroring

Client : java Win2k



We are expecting about 20'000 events per day (and probably more with the addition of automatic ticketing) for more than 100'000 supported users worldwide. We want a backup every 4 hours.



As I am neither an SC nor a storage specialist, here are my questions :

What should be our backup strategy ?

Should we use the 24x7 or stop mirroring on the SAN and make a cold backup on the mirror ?

Is there a way to do incremental backup with P4 ?

Is our backup frequency realistic (6x/day) ?



Thanks for your help.



Nicolas Chauvin

[tommy]

I can't answer all Your questions but I can tell you about our backup strategy.



We run SC3SP3 on mvs on P4.



We run 24x7 hot backup every second hour. The backup runs under normal conditions for about 7-8 minutes during which time users have slower response times.



We register approx 20.000 calls and approx 8.000 problems a month. We are also are 24x7 helpdesk with offices in USA, Denmark and Japan.



We have experienced hangups from time to time, it seems to occur if several events occur at the same time during the -stoplogging process. We have not yet been able to identify which events that cause these hangups. The result of the hangups is usually that we need to restart SC. Sometimes we are able to kill the deadlocks that are the result of the hangups.

[chauvin]

Thanks for your answer Tommy.

[eisefr]

Hi Tommy..

Did you find out what causes those hangups after starting the logging mode yet?

I just had the same problem too this morning.. :-/

[tommy]

No, I never found out exactly what triggers the problem.

I know this won't help You much but since we upgraded our server from SC5.0.1 to 5.0.1.7 we have not had any hangups after backup - knock on wood ;-)

[ambarnes]

I would simply suggest moving your incremental backups from every 4 hours to every 2 hours given that volume of traffic. Heck we don't have nearly the amount of volume you are going to have and we backup every 2 hours - but then we are also paranoid somewhat

[eisefr]

How is your backup done?
Just a copy of the data folder through a simple copy-command or are you using a software which saves the datafolder?

[tommy]

Yes basically.

We are on unix so we do a tar on the data folder to a new folder on a seperate disk system. After stoplogging we backup this folder to TSM.