Welcome to the SRP Forum! Please refer to the SRP Forum FAQ post if you have any questions regarding how the forum works.

RTI_Remaketable

My client has an issue with multiple GFE's on various really old, really big tables. Using Fix_LH, including the version in the Database Manager, returns an error immediately. If I use the RTI_Remaketable routine, how/will it react if/when it hits a GFE? I don't know the inner workings, but it seems like it's faster than doing record-copies between tables.

Thanks in advance!

Comments

  • It is my understanding that RTI_RemakeTable just does a standard read (from source) and write (to target). So there are unresolved GFEs then I would expect it to produce an error.
  • It does indeed!

    Do you know where Fix_LH will try to create the "dump table" when fixing? I'm getting the FS255 error, which I assume means it can't create said file. I'm thinking there are access rights at work.

    Thanks!
  • Either in DATAVOL or whatever is configured as the Default Data Path.
  • ok, I was able to Finally get Fix_LH to work. But due to network errors last night, group 0 of the table in question is corrupt now. Doing a Select returns a zero, even tho a Count returns a couple million records. I think the data is all there.

    The error on LH Verify is "Table size 1024, framesize , modulo ; table size/framesize <> modulo"

    How can I repair the header on the table? I've never done That before.

    Thanks in advance!
  • I don't have an answer to your direct question. How I normally handle these scenarios is to try and copy out as much data as possible while navigating around the GFEs. When I successfully copy out a row and remove it from the source table. Both the source and target tables are sizelocked to avoid issues with table resizing.
  • Tried that. With group 0 being hosed, I can't even Readnext through the table. Selects, Counts, etc. all return 0.

    I can try to restore from backup but the backup likely has a GFE too.
  • So a latent select also fails?
    Do you know the Key IDs? If so, you can write a loop that works with Key IDs that you generate rather than relying upon ReadNext.
  • I have A list of keys from a manual "index." Probably not complete but it's a start. Running now.

    Thanks!
  • Cool. Hopefully if you can skim off the majority of records this way, the remnant will be much easier to resolve if not fix.
  • The one table has been fixed and is fine. Now I have another that is Completely hosed, and it's constantly in use.

    Am I safe to use the GFE Fix option from Database Manager on an active table?

    Thanks!
Sign In or Register to comment.