execute migrate_content error [DM_STORAGE_E_RENAME_EXT_SEARCH]

hendra.liman

Hi all,

i need some help here. Currently we had a project of particular bank with massive document (6,000,000 objects ++ and still counting) and massive storage (about total of 5-6 TB with separated drives). They need us to do an Archiving procedure from one filestore (drive E:\) to another filestore (drive G:\). Im using this query:

Execute MIGRATE_CONTENT With target_store='filestore_archive_01', max_migrate_count=50000, remove_original=True, sysobject_query=True, log_file='C:\\LogArchiving.txt', type_to_query='ct_mandiri_doc(all)', query='r_creation_date < DATE(now) - 180 and a_storage_type <> ''filestore_archive_01'''

or in other word, currently we put everything on drive E:\ (filestore_01) and want to archive the old document (older that 180 days) to drive G:\ (filestore_archive_01)

in the first month, the this query runs normally, meaning drive E:\ gaining some free space and drive G:\ losing some free space as well which is correct. However, in the last 2 month until now, we are greatly losing space in drive E:\ (because of daily transaction activity) and drive G:\ has no changes, so i suspect the Archiving is not running correctly.

When i check the log, it says :

Mon Oct 29 23:26:15 2012 891000 [DM_STORAGE_E_RENAME_EXT_SEARCH]error:  "Unable to rename a file matching (e:\Documentum\data\MANDIRI_REPOSITORY\content_storage_02\00018a39\80\2e\36\42*) to (e:\Documentum\data\MANDIRI_REPOSITORY\content_storage_02\00018a39\80\2e\36\42.pdf). Operating System error: The system cannot find the file specified.

"

Mon Oct 29 23:26:15 2012 902000 [DM_CONTENT_E_MIGRATE_CONTENT_INVALID_CONTENT_FILE]error:  "The content file for dmr_content object 06018a3980596100 is invalid or not found or inaccessible."

Mon Oct 29 23:26:15 2012 927000 [DM_CONTENT_E_MIGRATE_CONTENT_ROLLBACK]error:  "Migration rolled back due to errors."

Mon Oct 29 23:26:15 2012 941000 [DM_CONTENT_T_MIGRATE_CONTENT_OBJECTS_MIGRATED]info: "0 object(s) successfully migrated."

Mon Oct 29 23:26:15 2012 941000 [DM_CONTENT_T_MIGRATE_CONTENT_TOTAL_SKIPPED]info:  "0 objects skipped during migration"

The bad news is, once i got this error, the system ROLLED BACK everything that we had already processed on that day. So during that execute migrate_content process, either it has to be 100% success or ROLLED BACK everything.

What i have tried:

1. run dmClean job

2. run dmFileScan job (took very-very long time because of massive amount of data)

3. Installation owner (Administrator, as well as admin user of windows) has Full Control of E:\Documentum\Data.

Question :

1. What is the cause of that error ?

2. Is it possible to BYPASS the invalid object access, and continue to the next valid object without rolling back everything ? Because its very2 waste of time.

Please.. somebody respond ASAP, because we only had 350GB on drive E:\ with transaction DAILY growth of 10GB.

Thanks before and after

Pingu22

Did you solve this problem? How? Please help

Andrea Paci

This KB article may be of your interest: https://support.emc.com/kb/410643

Make sure you follow best practices regarding cleanup/consistency jobs (see KB and above post).

What is the exact error that you are getting?

Did you try to dump/access the objects reported in the error message?

If you did, are you getting any problem with them? The content file may be missing/corrupt or the object in bad shape/corrupt...

Let us know.

Pingu22

Hi Andrea,

The error that I'm getting is the folllowing

[DM_STORAGE_E_RENAME_EXT_SEARCH]error:  "Unable to rename a file matching (e:\Documentum\data\<REPONAME>\content_storage_02\00018a39\80\2e\36\42*) to (e:\Documentum\data\<REPONAME>\content_storage_02\00018a39\80\2e\36\42.pdf). Operating System error: The system cannot find the file specified.

The object doesn't exists in the filesystem.

I already ran cleanup/consistency jobs and the error remains.

Somehow it's trying to access a document that doesn't exists in content's folder

matthohnke

1) You can try the batch size parameter for execute MIGRATE_CONTENT to avoid a large rollback.

EXECUTE MIGRATE_CONTENT

WITH TARGET_STORE = 'filestore_01',

...

BATCH_SIZE=200,

2) So you have the dmr_content record ... 06018a3980596100, so you can find the parent id on the dmr_content record - which gives you the dm_sysobject r_object_id.  Then just explicitly exclude it from your migration for now - by excluding it in your query.  

3) The cause is an inconsistency, in that there is a dmr_content record and no file on the filesystem, this should be in the Consistency job report.   Also since you found the r_object_id, try opening the document, given there is no filesystem file, this should fail also.

Andrea Paci

Hi,

I suppose that you mean that the content file referenced by the dm_sysobject / dmr_content does not exist in the file system. In that case you will need to restore that content file from backup. We cannot tell the reason why it has gone missing at this stage.

MIGRATE_CONTENT does not know that the content file is not present in the file system. It will try to access it as it is supposed to be there as there is a reference to it in a dm_sysobject / dmr_content combo.

You have to restore the content file from backup or delete the object is not necessary (chances are that it may be of an unrequired version after all...).

What is the full error message?

What objects / object types are linked to the error?

Did you try to dump/access the objects reported in the error message?

If you did, are you getting any problem with them? What exactly?

You ran clean up jobs (DMClean and DMFilescan). Did you run the resulting scripts?

You also ran consistency job (Consistency Checker). Did you fix all inconsistencies?