BE changes media set after using tape

[stauron]

OK, I have read several of the threads and most of the unhelpful, unorganized tech notes on Veritas' site and I think the answer is "I'm screwed" but just in case.

I am new to veritas, I have used legato in the past. I have been working on this project for about 3 weeks. We have several different jobs and most come out of the "media set 1" big bucket. We have a job that runs constantly (like every 30min) and we want to overwrite as necessary because we only need the most current for this job.

Every time BE gets done with the current backup of the job it puts the tape back in the wrong bin. It is happy to get them from the "tkcsdb" media set, but it dumps them back into the "media set 1" when it is done.

So my question is: will partitions solve this or am I missing something big or is there a better answer?

Thanks a ton!

 

[NassauBoy]

Greetings Stauron.
You haven't provided very much information in your post about the job or how you are using media sets, however I think I can point you in the right direction anyway.

Very importantly, the order in which the jobs get and use tapes to run is a very specific order and can trip you up right out of the box. It seems to me that you have created a media set (bin) called "tkcsdb" where these tapes for this recurring 30min job should belong. The other important information is the relationship between the media location called "scratch" media and the default media set called "media set 1".

I am assuming that the tkcsdb job is configured to use the tkcsdb media set (bin) to get tapes and overwrite the data there every 30 minutes and put the tape back into tkcsdb media set. If the job does this and then procedes to put the tape back in "media set 1" then I believe it didn't initially get the tape from the tkcsdb media set to begin with - it only appeared to do this, let me explain.

I think that even though the tkcsdb job is configured to get and overwrite the tape media from the tkcsdb media set, the overwrite and append settings on the tkcsdb media set won't allow it to. If this is the case then the job will follow a very specific order and look to the "scratch media" set location first for available tapes, and finding none there, will scan the other media sets for anything that it can find. Once it comes across the tkcsdb media set, it takes the tape and runs the job.

Because the job initially did not getm the tape from the tkcsdb media set, than it will not put it back there. It considers the tape a homeless tape and promptly stores in it's default media set home - media set 1.

Check in two places.
1. The properties of your tkcsdb media set.
2. The Backup Exec OPTIONS on how BE jobs find and use tapes when it needs media or it needs more media to finish jobs.

I hope this helps and if not, it will get you to look a little deeper into the veritas backup procedures and then point you to the correct solution.

Good luck.

 

[stauron]

Nassauboy,

Thanks for your suggestions. I think you are on to something. I have checked the media bin countless times and it is set up the way we want.

BUT, in options on the Media Overwrite tab, the Media overwrite options are scratch first and then recyclable.

The tkcsdb Media Set is 0 hours overwrite and Infinite allow append, Media set one is overwrite 33 days and infinite append. I put the tapes in the tkcsdb, they are used and wind up back in media set 1.

So if I switch to the recyclable first then scratch, it seems like it should work.

I will give that a try.





Jeff Halfhide
Duvoisin & Associates

 

[justin2000]

stauron

if you want backup exec to put the tape back into the tkcsdb media set, you have to tell the backup job to put it in there. You have it currently set to put the tape into Media Set 1 once completed.

The way you have setup your media sets means this..

tkcsdb - if the backup job is set for overwrite, it will use the same tape as the last backup after 0 hours.(thats if you have backup exec set to use recycable media first, then scratch media.
If the backup job is set for append to tape, if you can't find one, overwrite it - it will append to the tape continuously until the tape is full.
If the backup job is set to append only, then terminate - if will append to the tape until it is full.

Media Set 1 - If the backup job is set for overwrite, it will use the tape, only if the last time it used the tape is 33 days and 1 second ago.
If the backup job is set to append to tape, if you can't find one, overwrite the tape - it will append to the tape continuously until it is full - then the overwrite protection period kicks in and you can overwrite the tape after 33 days and 1 second.
If the backup job is set to append, then terminate, it will continuously append to the tape until full. It will never overwrite the tape in this type of job.

Hope this explains your media set further.

 

[stauron]

Rats, I was just thinking I had the hang of this. So where do I tell the job to put the tapes back in the tkcsdb bucket when it is done?

On the Job Properties page there are 4 fields and all of them are set to tkcsdb. Job name, Media name, Backup set description, Media set. Justin2000 is this the place you are talking about?

Jeff Halfhide
Duvoisin & Associates

 

[justin2000]

stauron, thats right the Media Set name is where it will put the tape once the backup is done.

 

[stauron]

Well the Media set name is tkcsdb and they still end up in media set one when the job is done.

This is really the problem. I don't want to have to move them back everyday.

Is there another option that controls that?



Jeff Halfhide
Duvoisin & Associates

 

[jbud]

While I can't verify it with any documentation, I think the problem lies with the fact that your overwrite protection on the media set is set to 0.

I would suggest you set it to 1 hour, and since you're backing up every 30 minutes, use 2 pieces of media. In 8.6, I can't set the overwrite protection to fractions of an hour.

 

[stauron]

Still the same. It goes back every time to the default...

Will the partitions help this? Anyone?

Jeff Halfhide
Duvoisin & Associates

 

[jbud]

I suspect you have something internally wrong with the database structure. I would try (if you haven't already), creating a new media set with a different name and using it. What happens if there is no available media in the tkcsdb set, but scratch media is available?

What version are you running?

Very interesting...

 

[stauron]

Ok, thanks for your help so far. Here is the interesting bit. I am running BE 8.60 rev 3878. I watched the tkcsdb job run. I set the overwrite time on the tkcsdb job to 1 hour. I have not yet created a new bin for the job.

It waited for me to take the tape from media set 1 and put it in the tkcsdb media set. Ran the job and then dumped the tape back in media set 1. The really wild thing is that the tapes described as tkcsdb come into the bin with overwrite protection time of 33 days, which is the media bin time, but the append time is set for infinite. They are the only tapes in the bin set for infinite. The other jobs get put back in with the settings for media set 1.

It either hates me or is possessed.

Jeff Halfhide
Duvoisin & Associates

 

[stauron]

OK, OK I think I figured it out. I believe that the problem was with the name of the media bin. I listed it as only tkcsdb because I didn't want to type out the whole name. The full name was longer and had a hyphen. I created a new bin without the hyphen and it is working now.

Thankfully I discovered that now, as I am in the process of redesigning the media bins, 2 of the others had hyphens too.

Thanks for everyone's input. Sorry for being a spazz.

Jeff Halfhide
Duvoisin & Associates

 

[Qdog]

Your backup job is targeting Media Set 1 or it would not put the tapes in that media set when the job finishes. Change the backup job properties to target the media set you want and change the media set overwrite protection to something other than Infinite - Do not allow ovewrite.

 

[stauron]

Qdog,

Thanks for your input but I am sure the job targets the tkcsdb media bin. I have checked the properties multiple times.

Currently the job is set to: Append to media, overwrite if no appendable media is available.

Destination media set: tkcsdb

The problem is that something misfires. Originally (when I first submitted the post) several times per week the tapes would all end up in the media set 1 bin, the job would fail and ask for media. The job pointed to the media set we wanted, but the media set was empty.

Since then I have created a new media set and pointed the job to the new media set. That worked somewhat. It now runs for about a week and then dumps the tapes into media set one and not tkcsdb.

I manually move them from media set 1 to tkcsdb and it runs that day while I watch fine. The job runs every 1/2 hour, so I can watch it closely. I have 3 tapes in the media set and the media set is set for 1 hour overwrite protection and infinite append.

So I understand why you would think that was the problem. BUT IT AIN'T. There is some other problem. I am happy to give more info so ask if I am leaving something relevant out it is because I don't know.

Thanks again for all your help.

Jeff Halfhide
Duvoisin & Associates

 

[stauron]

Ok, I am still not getting this. I have several media bins Full, Media Set 1, tkcsdb. The jobs randomly move the tapes around.

My real goal is to have the tkcsdb be a separate tape because we only need the most recent backup and then it can overwrite. I also need to have the weekly fulls on 3-4 tapes so we can move them offsite. Right now they are just mixed in on like 8 or so tapes.

I have checked the properties multiple times, and the seem to be the logical ones.

Please someone help. Tell me what other details you need.

Thanks

Jeff Halfhide
Duvoisin & Associates