ndsrepair error 69/70 on eDir 22.214.171.124
Midst IDM upgrade here from eDir 126.96.36.199 / IDM 4.6 to eDir 9.1.3 / IDM 4.7 and seeing something strange that I can't find any real information on. Hoping maybe somebody else can shed some additional light on it.
The setup is currently a six server tree. Three eDir9 (new) and three eDir8 (old). We're migrating from old to new, and being mid-migration, I need all six working for a while. There is only one partition, and eDir9 holds the master. The other five servers are all read/write. There is an IDM 47 engine on eDir9 doing *most* of the work at this point, and an IDM 46 engine on eDir8 running an eDir-to-eDir driver that has not yet been migrated to the eDir9/IDM47 engine.
For other reasons (Support SR), I needed an 'ndsrepair -R' on all six servers in this tree. On the three eDir8 servers, I got a shed-load of error 69 / error 70 like:
Error found: 69 General => Locale: 6, Logical File: 0ae2, Logical File Type: 00 Block => BTree Level: 00000000, Block Address: Record => Elem Offset: 0000075e, DatRec Number: 00008023, ElmRec Offset: 0000075e, Field Number: 00000000
The details change with each of the thousands of reports, but the basic format is the same. A second run of 'ndsrepair -R' was clean with no errors. That was a couple of weeks ago.
I have a second tree, virtually identical to the first. Built the same way. Migrated to the same state. And, lucky me, the same issue that prompted the SR in the first tree, also happened in the second tree. And I got to have a second session with ndsdump, followed by a 'ndsrepair -R' on all six servers. And, again, the eDir8 servers all reported a shed-load of these error 69 / error 70 errors. That was Friday.
This morning, going back to the first tree, 'ndsrepair -R' reports a shed-load of error 69 / error 70 when run on the eDir8 servers, and no problems on the eDir9s.
Re-checking the second tree, 'ndsrepair -R' this morning (which was clean on Friday), same thing. Lots of error 69 / error 70 errors reported.
The only thing I've found on error 69 / error 70 is that they are FLAIM level errors. Oh, and it's not the same as this thread:
at least not as far as I can see, because that one is about errors on eDir9, which I'm not seeing.
My intent is to decomission the eDir8 servers once the migration is complete, but that won't be for another few months. Meanwhile, should I be concerned about these errors? I would normally have assumed that the answer would be yes, but the tree seems to be otherwise working fine. I don't think that they're cosmetic errors, but they don't seem to be hurting anything either.
Would you worry?