RMAN Space Reclaim

Dear Gurus,
I'm using RMAN with third party application for backup to disk. Our policy is to keep daily backups for 1 month and monthly backups for 1 year. I'm not setting RMAN retention policy. I used set until in rman scripts for setting the daily and monthly retention periods. However, I can not delete obsolete/expired backups because I'm not setting the retention policy.
This is sample of my script:
RMAN> backup
keep until time "sysdate+30" logs (database);
RMAN> delete obsolete;
RMAN-00571: ===========================================================
RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS ===============
RMAN-00571: ===========================================================
RMAN-03002: failure of delete command at 08/29/2009 16:03:35
RMAN-06525: RMAN retention policy is set to none
Please advise how can I reclaim space of expired backups.

HI..
From metalink:--
Error:     RMAN-06525 (RMAN-6525)Text:     RMAN retention policy is set to none
Cause:     Command DELETE OBSOLETE and REPORT OBSOLETE requires that either:
     * RMAN retention policy is not NONE or, * RMAN retention policy is
     specified with REPORT/DELETE command.
Action:     Either configure RMAN retention policy with CONFIGURE command or
     specify it at the end of DELETE/REPORT command.
>
Refer to [http://download.oracle.com/docs/cd/B19306_01/backup.102/b14194/rcmsynta040.htm#i83923]
Anand

Similar Messages

Space reclaimer on Thick Provisioned LUN

Hi, Could someone clarify something for me. I understand the need to reclaim space on thin provisioned LUNs using SnapDrive's space reclaimer. Is this a requirement for thick provisioned LUNs or will running it only fix the space reporting in System Manager. As an example - a 10Tb thick provisoned LUN is presented to a physical Windows server and mapped to a drive. 6tb of data is written to the LUN, and Windows and System Manager both agree that there is 4Tb free. All of the data is deleted. Windows sees 10Tb free and System Manager still sees 4Tb free as it has no way of knowing the blocks have been marked as deleted. Expected behaviour. Now if we write another 6Tb to the volume will there be space? The LUN will eventualy show 100% full, but after this will Windows quite happily carry on writing to the drive? I could test this but I wondered if anyone had any real world experience. In my organisation we have a process regularly run Snapdrive space reclaimer on thick provisioned LUNs and I am wondering if we need to (other than getting system manager to report correctly)? Many Thanks, Jason

It's still complicated I do not really know what combination of settings System Manager is using for "thick provisioned"; if we speak about space gurantee == volume and fractional reserve == 100% then you should always have enough space to (over-)write full LUN size.But note that under some conditions NetApp may not honor volume space gurantee ...

Space Reclaim

Dear Experts,
I need to reclaim space from huge indexes(Segment advisor showing lots fragmentation). What is the best way to return that fragmented space to tablespace. We also have space constraint i.e. size of index is so large that we don't have enough free space as the size of index. :)
Thanks&Regards
Sunil Kumar

sunil kumar wrote:
I have not even free space to rebuild the index, as it require at least size of the index free space for rebuild, as I mentioned earlier we have huge index to rebuild i.e. around 200GB, So what could be the better option !!.
Not only do you need free space to put the rebuilt index, you will need a large amount of TEMPORARY tablespace to write the intermediate sort runs.
Can you give us some idea of the scale of the problem - you've said that there is a "lot" of wasted space, but that doesn't mean much. Can you give some examples of: (current size, predicted size). Basically I wouldn't be in a great hurry to rebuild a large index unless the predicted size was less than 50% of the current size. I'd take note of it, of course, and ask myself why the index had got to that size - the answer makes a big difference to what you do about it.
If you're sure you need to rebuild lots of indexes to reclaim space, and you don't have the space to do it - what about starting with the smallest first, as this may gradually free up space that can be used for the larger rebuilds.
Regards
Jonathan Lewis

Table Partitions and Rolling Window: free space not reclaimed

I have a couple of processes that involve the rolling window scenario where we drop an old partition and add a new partition monthly. We are seeing, however, that free space is not being reclaimed, and so I keep having to feed new datafiles into the tablespace (which is LMT). I was (mistakenly, it seems) under the impression that dropping a partition frees up that space for re-use.
We're currently on Oracle 10gR2 (10.2.0.2), but these tablespaces were orginally created in 9i. I've read that one solution is to implement "table shrinking," however I also see that is only available on tables in an ASSM tablespace. I see that ASSM is the default for tablespaces in 10gR2, but since my tablespaces were created in 9i, I assume they are not ASSM.
I'd like to see what other people think of this scenario. Surely a lot of people do the rolling window scenario and need space reclaimed.

No, it is of concern because Sanadra was messing with the partitions. She could have inadvertantly caused something to happen, in this case the free-space issue, and it's imperative to get all of the details.
My iMac 27' Late-2012 does not have a fusion drive, yet shows that the type is "internal" the original poster doesn't indicate that.

NOT ABLE TO RECLAIM STORAGE USED BY XMLTYPE COLUMN

Since we are on Oracle 9i, the 10G solution DBMS_XMLSCHEMA.CopyEvolve() is not available, we are trying to do it by ourselves.
see the doc,
Re: how to make xml schema change when there are existing records in the ta
However, our requirement is different in the following ways:
1. We hope to be able to add/drop as many XMLTYPE columns to a table as we
want( all structured storage, but with different schemas).
2. When a XMLTYPE Schema needs to be updated, we will add a new
xmltype column, assign it the new schema and migrate the data from the old column. After that we want to drop the old xmltype column and it's schema to reclaim all the storage.
The underlying xml tables are drop, space reclaimed.
However, we found out that the segment/blocks/bytes of the xmltype column itself is not released.
Question:
1. Since Xmltype column associated with xmltype schema is structured
stored, the xmldoc should parsed and stored in underlying tables. However, it seems that the xmltype column itself is acquiring sunstantial amount of space.
By querying user_segments table, we saw the underlying structure is actually taking constant and small space, however, the main table is taking huge space.(We populated the table with 61K rows), For instance,
SEGMENT_NAME BYTES SEGMENT_TYPE
PERFORMANCEEVENT 260046848 TABLE
SYS_C007080275 1048576 INDEX
Performance1833_TAB 65536 TABLE
Date1831_TAB 65536 TABLE
LocationAddress1814_TAB 65536 TABLE
SYS_IL0010401244C00018$$ 65536 LOBINDEX
SYS_IL0010401244C00012$$ 65536 LOBINDEX
Where PERFORMANCEEVENT is the main table.
Why?

Try to read this post.
This will give you an idea about the way and when xml is shred into tables.
XMLType column based on XML Schema: several questions
In your case since you are not using the default table, the data is stored witin the table itself as oracle has not shred the data.

Running out of disk space (large VM boxes)

Gentlemen:
CUCM 6.1.4
Unity version = 5.0(1.0)
Cisco Message Store Manager Version 1.4.0.38
Exchange Information Store Version: 6.5.7638.1
Running out of disk space (large VM boxes)
Here is the situation: My client never set up any storage or flush procedures within Cisco Unity Tools Message Store Manager (MSM). Now they have several voice mailboxes with a half a GB of data or more in their (Deleted Items Folder)
I ran the (flush messages from deleted items folder) with success on these voice mailboxes.
The question is should I see some disk space reclaimed on the Server Hard Drives?
Thanks, Tom

With Exchange, you will not see disk space reclaimed until an offline defrag is run against the Exchange databases.
http://support.microsoft.com/kb/328804
Hope this helps.
Brandon

Cannot recover hard disk space when deleting and huge sparseimage

hi there,
System Version: Mac OS X 10.4.3 (8F46)
Kernel Version: Darwin 8.3.0
my problem is that I found the mac was running slow and I noticed the harddisks were getting full:
[veronique:/Users/e] e# df -H
Filesystem Size Used Avail Capacity Mounted on
/dev/disk0s3 60G 59G 729M 99% /
devfs 102K 102K 0B 100% /dev
fdesc 1.0K 1.0K 0B 100% /dev
<volfs> 524K 524K 0B 100% /.vol
/dev/disk1s2 60G 59G 725M 99% /Users/e
automount -nsl [246] 0B 0B 0B 100% /Network
automount -fstab [250] 0B 0B 0B 100% /automount/Servers
automount -static [250] 0B 0B 0B 100% /automount/static
Im not sure why they became this full but.. I then looked for big files and found a sparseimage of 40GB. I can mount it and all so it s in good shape.
I needed to free up space and I delete a lot of music files, pictures, a virtual pc file of several gigs and so forth, but when I ran a new df -H things had nt changed.. So im guessing all the data, inclusive deleted data is in the sparseimage? is this so? I ve emptied the trash several times.
I ve now tried to turn filevault OFF but it asked me to free up 1.9Gb.. catch 22.. if I rid myself of files it wont see more freed up space...
is there something I can do? I could mount the sparseimage and browse it and delete files there or back up whatever I need and then get rid of it? should I re install?
any help needs to go to [email protected] i went into my settings but didnt find out how to change my email. thats not impressive i know
eric

Here's what I had to resort to to forcibly reclaim missing free space on my FileVault-encrypted PowerBook G4.
After several days and more restart/shutdown trials than I can count, I could not for the life of me get my PowerBook to give me the "Reclaim free space" option that usually pops up. Not sure why. But, I was missing nearly 12GB of free space and kind of needed it.
I was able to force a space reclaim... but the process isn't exactly point-and-click.
•I mounted the PowerBook via Firewire target disk mode on my desktop and copied over my PB account's .sparseimage file to my desktop (didn't want to work on the original file in case I screwed it up).
•Used the terminal command "hdiutil compact" on the .sparseimage file (type "hdiutil compact " (with a trailing space) then before hitting return, drag the icon of the .sparseimage file on to the terminal. That'll insert the path of the file. Hit return.)
•Let it run. Took about 15 minutes or so. On a 27GB .sparseimage, the procedure recovered 12.5GB!
•Deleted the original .sparseimage from my PowerBook, emptied the trash, and copied over the newly compacted one. Rebooted the PB, and voila!
Do keep in mind though that you'll need at least as much free space as your Home folder occupies itself if you want to turn FileVault off.
PowerMac Dual 2.5GHz G5 Mac OS X (10.4)

OCFS / CRS / IA64 / RHEL 3 / 10g Support?

As I described in another thread, I have the following products and CRS does Not detect nor allow me to use the OCFS Filesystem when I try to input the device name/filesystem name, it errors and fails in the installer screen. The Support site for Oracle is Vague on this configuration officially. OCFS is UP and Running on four nodes. I actually installed Enterprise 10G on one system and defined the OCFS partition for my DB files. Worked fine. CRS seems to be my problem...
ocfs-2.4.21-EL-1.0.13-1.ia64
Oracle Cluster Ready Services Release 1 (10.1.0.3) for Linux Itanium
[root@sysl012 root]# uname -a
Linux sysl012 2.4.21-15.EL #1 SMP Thu Apr 22 00:13:07 EDT 2004 ia64 ia64 ia64 GNU/Linux
ANY official "support" for this?

Hi,
ORA-19816: WARNING: Files may exist in db_recovery_file_dest that are not known to database.This looks like an ASM-full condition, see here:
http://dba.ipbhost.com/index.php?showtopic=11393&mode=linear
ORA-19816: WARNING: Files may exist in string that are not known to database.
Cause: One of the following events caused this:
1. A database crash happened during file creation.
2. A backup control file was restored.
3. The control file was re-created.
4. DB_RECOVERY_FILE_DEST has previously been enabled and then disabled.
Action: Use RMAN command CATALOG RECOVERY AREA to re-catalog any such files. If the file header is corrupted, then delete those files using an OS utility. Do not use messages 19817; it is used for simulating lock failure Do not use messages 19818; it is used for space reclaimation before backup
Hope this helps. . .
Donald K. Burleson
Oracle Press author
Author of "Oracle Tuning: The Definitive Reference":
http://www.dba-oracle.com/bp/s_oracle_tuning_book.htm

Index rebuild

Hi Team,
Please tell me any difference b/w alter index <index> rebuild or alter index <index> rebuild move.
Is there any difference that more space will be released if I move index to diffrenet tablespace and move move back to orginal tablespace.
I know that if we use move option we can move index to different tablespace,
Regards,
jit

Even if you would do lots of dml's, that's not certainly means that you need the rebuilding of the index. If your assertion is based on this that there would be some "holes" within the leaf blocks which wont be reused and would require a rebuild, it's wrong then since Oracle would take care of these things on its own. So again, what criteria, other than an OLTP db you used to conclude that you need rebuilding of indexes?
But I just wanted to know If I go for alter index rebuild or alter index rebuild move and move back the indexes, which option will be best for me mean by which option I can reclaim more space.Well, I am sorry to say but I think you really don't know very well both the commands. The commands don't differ in the space reclaim given by them but in the way they do the rebuild. Please read the documentation to know what both commands do and how? BTW, from where you picked the command alter index rebuild move since I am not sure that such command exists?
Aman....

Best way to remove old data from db?

Hi,
I am looking some ways to keep my database small and remove old data from database and archive it somewhere else. the data may need to be put back to the database in the future. One of the things I am conserned is how to enforce constraints? In other words, when I remove data, all the related pk/fk data should be removed together. Any suggestions will be greatly appreciated.
Thanks

hope this may help u
concept is : exp/imp & truncate will reclaim the space
method1
do an export and secure the dump(say dump1) for future purpose.
delete the unwanted data.
commit.
now again export the tables ( say dump2).
now do the import(dump2).
/*now this import will reclaim the the unused space.*/
method 2
/* depennded on volume of data */
-- create backup of existing table for future
create table backup_table nologging as select * from original_table;
-- create copy1 of existing table
create table copy_table nologging as select * from original_table;
-- truncate the original table, so that highwater mark is reset.
truncate table original_table;
-- insert data from copy_table into original table
insert /*+ APPEND */ into original_table select * copy_table where(ur conditions for data);
commit all the work.
/* now u have 3 tables
backup_table - with all the data - for future reload
original_table - with unwanted data deleted and space reclaimed.
copy_table - copy of original table.
now u drop the copy_table
drop table copy_table;
regards

? - Can I Test Mavs and Back-up All Changes to SL - ?

SL is great and I really don't want to leave.
Having said that, it's probably time to at least do a good hard test on Mavs before I'm so far out of date that I'm essentially teletype circa 1900.
So, I'm hardware good and application good (installed Mavs on a bootable external and can tell that from a quick run on 3 different iMacs at work --- but running off of an external via Firewire is too slow for extended work and/or purposes - just did that as a 1st round check on things).
So here's my question -
If I partition my HD; then SuperDuper my existing 10.6.8 to the new partition, install Mavs on the new partition, boot from the Mav's partition, use that for a few days to see how things are going, and assuming as I do all will be well --
IF all is not good with Mavs, and IF I want to revert back to SL in a week or two after running Mavs exclusively, since I am in a small business situation and need to keep everything up to date, is there any way to keep the SL partition up to date as I go along - so if, for example, I amend a contract, or create a new folder, or do whatever on the Mavs partition, to have it "mirror" to the SL partition?
More to the point - Any way to "sync" my 2 partitions so that the SL partition stays current with the Mavs partition? (I use that term loosely as I am a professional with restrictions on what I can "cloud" store).

The only issue with that arrangement is that some Mavericks versions of the applications, i.e., Mail, iPhoto, Aperture, will ireversably modify the data files so that they won't be able to be used by the Snow Leopard versions of the apps.
I agree that running from an external partition doesn't normally make a big performance difference, the exception might be an older USB1 drive.
One can easily add a partition to their internal drive and install say Mavericks on that addition. If you decide against using Mavericks, the partition can easily be removed and the space reclaimed for the original SL partition.
If he decides that he no longer want the SL partition, then it would be little more complex and time consuming to arrange. He would have to erase the SL partition, clone the the Mavericks partition to the SL's prior partition, then remove the outer Mavericks partition and expand the inner Mavericks partition to reclaim all of the space.

What's with Sparseimage Size?

I currently have a sparseimage in my Macintosh HD that has the contents of a corrupted user ID. I have changed the permissions and I can get to files and have been moving off the photos and music files to an external hard drive.
I thought that after moving and deleting those files in the sparseimage, I would see a jump in available space on my Mactintosh HD (these files came to about 8 GB), but even after emptying my trash, the available space has remained the same. What is going on? I'm thinking the sparseimage size was determined at the time it was created and that removing files does not change the amount of space it takes up. Is this so?
Once I move the files I want to keep, I will then have the option of deleting the sparseimage altogether. Will this allow me to recover that space?
-Bill
iBook G-4 Mac OS X (10.4.7) 1.33 GHz, 768 MB memory, 60 GB HD, iPod Shuffle

Hi mark,
Your comment was very insightful. I suspect the reason the sparseimage is not shrinking is that I am unable to access it through the user ID with which it was created. I do recall that when I would log out in the past, File Vault would ask if I wanted to recover the space, which I usually did. Now that I am unable to do that due to the corrupted ID, even though the files have been removed, the space remains the same. One ancillary question: If I am not using File Vault, is that space reclaimed as soon as the file is deleted and trash emptied?
I will use the hdiutil you mentioned until I am prepared to delete the sparseimage altogether.
-Bill
iBook G-4 Mac OS X (10.4.8) 1.33 GHz, 768 MB memory, 60 GB HD, iPod Shuffle

Estimation of archive generation

Hi Experts,
I am planning to shrink a segment in production db to reclaim the space.The size of the table is 217 gb and on calculation it's size estimated to be max 50 gb keeping the consideration of overhead in the blocks due to PCTFREE etc..The agenda is to estimate the archive generation so that I could be ready for the space at OS to accomodate the archive generation for this shrink activity.Pls suggest me for some way to estimate archives generation before this operation so as to make it successful..Here are the details that may be required by you
DB: 10.2.0.4 Enterprise edition
OS: server 2003 (windows)
index on table: 1 unique index of size 12gb

Hi Aman/Kuljeet
I have already tested this on t&d server but since the data and the size was quite small than real so came up with this question.However I observed that aprox 4 times the archives was generated than the space reclaimed..Now using this observation it comes to 600 gb in production aprox which was a concern for keeping our dataguard in sync..Anyways thanks a lot for your response.
Regards
Asif Khan

ASM disk resize

Hi All
I have a requirement to resize ASM disks to a smaller size
Currently My ASM disk group is have a single disk with 500GB
I want to resize it to 200 GB .The current disk usage is 50GB
How can i do it?
* Can an online resize of disk from OS level will help to resize the ASM Disk
* Does this operation is supported by ASM instance to perform as online
* Does the space reclaim will happen online without restarting asm instance
Additional info
rdbms: 11g R2 11.2.0.3.0
Asm redundancy : external

user7993571 wrote:
Hi All
I have a requirement to resize
Nobody here prevents you from doing as required.
Please proceed to do so without further delay.
How do I ask a question on the forums?
https://forums.oracle.com/message/9362002#9362002

[Slvd] What to restart after system upgrade? When to restart X server?

Hi all,
After a system upgrade, what should be restarted? The following actions I know about:
kernel (linux): reboot
daemon (sshd, cups, iptables, etc.): rc.d restart ...
application (fvwm, mutt, etc.): restart application
But what will require restarting X server? Can the current X server continue to run after upgrading one the following:
1. xorg-server
2. libx11 or other libraries xorg-server depends on
3. anything else?
I want to keep the system up-to-date but avoid restarting the X server to maintain my X session (>100 windows carfully spread and grouped over 22 workspaces). Unfortunately, there is no xsm anymore and, no, I don't like gnome/kde/thelike.
Thanks,
Daan
Last edited by drrossum (2012-03-26 15:37:37)

brebs wrote:
A useful command is:
lsof | grep 'DEL.*lib'
I put this line in a script called stalelibs. Helps to see which services should be restarted. Or whether I'll just take the easy option and reboot :)
This is very helpful!
Also, in the link that you sent it is explained why programs don't break when their shared libraries are removed:
Existing programs (that are already loaded) have pointers to the old version, as Naib says, in RAM as code or an inode location as a file handle to the deleted version. Once all handles are gone/programs terminated, the file will then truly be "removed" from the disk and space reclaimed. Until then you'll see shared libraries still eating disk space and RAM, despite them being removed.
This is why it's good to reboot the machine after certain upgrades such as libc, so programs like init can pick up a new copy of libc.
@brebs: Thanks for the great contribution!

RMAN Space Reclaim

Similar Messages

Maybe you are looking for