I personally like the idea of being able to resolve further on down the road. Can potentially save a step in troubleshooting.
Scott Klassen -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Brett Shirley Sent: Wednesday, February 22, 2006 1:35 AM To: [email protected] Cc: Send - AD mailing list Subject: RE: [ActiveDir] repadmin info oddity I wouldn't make a "hard feature" on such parsing, b/c it is not documented and could change at Brett or (lets call him) Greg's willy nilly. Oh and I just rememberd ... I should also note, isn't a 100% accurate to use retired DSA signatures, we punted a bug in repadmin relating around this, IIRC ... IFM maintains the same retired DSAs list as the original source DC, meaning it is tricky to figure who is the real slim shady when deal with retired IIDs. It may be probabalistically or even deterministically possible by looking at when the IID was retired, and matching against active IIDs, and matching retired IIDs, to figure out who originally owned the IID. That's what joe should work on... We've also debated when we delete a DSA or stamping all related IIDs, into a never cleared place in the DS, so that repadmin /showutdvec could always resolve DSAs even after they've been deleted for 10 years. On Tue, 21 Feb 2006, Dean Wells wrote: > Hmmm, I would guess he's probably adding a new switch to deal with this > particular thread. > > Anyway, since he's not responded, I'll take a stab at what ADfind can or > cannot do here (not really ADfind's problem if my lazy research is > accurate). The attribute in question's syntax is a single-valued "octet > string" which can typically be filtered against assuming the correct > notation is supplied. This particular attribute, however, will often > contain multiple GUIDs within the flat value (a pack of them) making it > difficult to successfully construct a reliable and/or optimal filter > (remember, medial queries are painful without the necessary index). > > To further complicate the issue, the byte ordering is maintained differently > internally to the way it's displayed. Since ADfind AFAIK cannot yet decode > "retiredReplDSASignatures", in order to query against it we have to reorder > it ourselves. Here's an example of how to convert repadmin's display format > to the internally maintained byte ordering (this is a little painful) - > > repadmin's output = 6cc4a8e0-2019-4e4f-81cd-f35926de38a3 > internal structure = E0 A8 C4 6C 19 20 4F 4E 81 CD F3 59 26 DE 38 A3 > > ... now trim the hyphens and pad repadmin's output to pair up the bytes - > > repadmin's output = 6c c4 a8 e0 20 19 4e 4f-81 cd f3 59 26 de 38 a3 (padded > & trimmed) > internal structure = E0 A8 C4 6C 19 20 4F 4E 81 CD F3 59 26 DE 38 A3 > > ... now, re-order the 1st 4 octets, then the next 2 octets and again the > next 2 octets. I've added extra spaces for legibility (essentially, you're > re-ordering the first double-word, the next word, the next word and the rest > remains as is ... this is known as "network" or "pretty" byte ordering) - > > repadmin's output = e0 a8 c4 6c 19 20 4f 4e 81 cd f3 59 26 de 38 > a3 (re-ordered) > internal structure = E0 A8 C4 6C 19 20 4F 4E 81 CD F3 59 26 DE 38 > A3 > > OK, having done all of that, you now have two possible options: > > option 1) use a fairly concise query and parse the output as follows ... > > ... create a string of 8 words (or 8 octet pairs if you prefer) to match > ADfind's output format - > > resulting structure = E0A8 C46C 1920 4F4E 81CD F359 26DE 38A3 > > ... then use the following syntax - > > C:\>adfind -config -f > "&(objectcategory=ntdsdsa)(retiredReplDSASignatures=*)" -csv -nocsvheader > retiredReplDSASignatures | findstr "E0A8 C46C 1920 4F4E 81CD F359 26DE 38A3" > > ... this returns the DN of the "NTDS Settings" object of the DC that owns > the retired invocation ID. If no results are returned, one of two things > occurred; 1) you fat-fingered it or 2) the DC no longer exists. > > option 2) submit the following v. expensive query (note, it's a medial > query) - > > C:\>adfind -config -f > "retiredReplDSASignatures=*\E0\A8\C4\6C\19\20\4F\4E\81\CD\F3\59\26\DE\38\A3* > " retiredReplDSASignatures > > Fingers crossed that Joe will have a hidden switch to do the decoding for > you, until then, this is it I'm afraid. > > -- > Dean Wells > MSEtechnology > * Email: [EMAIL PROTECTED] > http://msetechnology.com > > > -----Original Message----- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] On Behalf Of Scott Klassen > Sent: Tuesday, February 21, 2006 5:11 PM > To: [email protected] > Subject: RE: [ActiveDir] repadmin info oddity > > Too bad Joe picked today to be MIA. > > Scott Klassen > > -----Original Message----- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] On Behalf Of Dean Wells > Sent: Tuesday, February 21, 2006 7:59 AM > To: Send - AD mailing list > Subject: RE: [ActiveDir] repadmin info oddity > > The GUIDs returned in this scenario are not used by the directory in the > traditional manner and, as such, using a GUID-based binding string won't > locate the owning object. The invocation IDs (which are indeed GUIDs but > not objectGUIDs) are maintained on the DC's NTDSDSA instance (its NTDS > Settings object) by the "invocationId" property ... retired invocation IDs > are maintained by retiredReplDSASignatures. > > ADfind can likely hit these ... but the GUIDS needs to be expressed as part > of the query filter, not the base. I'll leave joe (why does he insist on > using a little "j"?) to provide the ADfind syntax (it seems that no matter > how hard I try, joe will always have a better switch ... and if he doesn't > have one, I'm fairly certain he quickly adds it, quietly releases the new > binary with the same version number, posts the reply, ridicules my > uneducated attempt at using a "real tool" (joe's words, not mine) and > professes his innocence :0). > > > -- > Dean Wells > MSEtechnology > * Email: [EMAIL PROTECTED] > http://msetechnology.com > > > -----Original Message----- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] On Behalf Of Thommes, Michael M. > Sent: Tuesday, February 21, 2006 7:44 AM > To: [email protected] > Subject: RE: [ActiveDir] repadmin info oddity > > Adfind (http://www.joeware.net/win/free/tools/adfind.htm) to the rescue! > I recently had to do this and got it accomplished with the following syntax > (with a little help from joe :) ): > > adfind -default -binenc -f > objectGUID={{GUID:0B3F5BC4-5713-4611-8F6A-752A3B0DE664}} dn > > ("adfind /???" For lots of good info!) > > Mike Thommes > > -----Original Message----- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] On Behalf Of SCOTT KLASSEN > Sent: Monday, February 20, 2006 8:56 PM > To: [email protected] > Subject: [ActiveDir] repadmin info oddity > > I try to keep up on new or updated MS KB articles and often check to see how > they correlate with my environment. I noticed that 875495, dealing with USN > rollbacks, was updated earlier this month. As I've experienced two AD > issues, both of which needed PSS involvement (one dealing with sysvol > inconsistency and the other which wound up being the RID master going on > > temporary strike) I figured that I'd do a quick check as described in the > article. On the good side, the USN's are consistent between controllers. > On the disconcerting side, I got a little more information than I was > expecting. Besides my DC's, I also got USN listings for several GUIDs. > I > assume these are leftovers from DC demotions and only remain in the form of > historical data. Do I need to worry about these (especially the DC1 > (retired) listing) and is there a way I can resolve the GUIDs to names, find > where this info is hiding, and clear them out? > > Thanks, > > Scott Klassen > > >repadmin /showutdvec dc1 dc=domain,dc=com Caching GUIDs. > .. > Default-First-Site-Name\DC2 @ USN 455091 @ Time 2006-02-20 > 20:08:20 > 2c92760e-e8fc-4418-947e-3b1016ab8514 @ USN 1012381 @ Time 2005-08-04 > 00:02:34 > 6e129965-56c3-469e-b70a-f1fdfb8bb2cc @ USN 969931 @ Time 2004-07-24 > 11:53:16 > Default-First-Site-Name\DC1 @ USN 1717571 @ Time 2006-02-20 > 20:10:50 > Default-First-Site-Name\DC1 (retired) @ USN 1298674 @ Time 2005-08-05 > 06:36:16 > e2199f22-f1dd-4d1c-90a6-0e8bb874f355 @ USN 744173 @ Time 2004-12-28 > 20:52:04 > ff0d7d50-214f-4bc1-96b6-55ac6ef317f0 @ USN 852323 @ Time 2005-06-08 > 14:29:20 > > > List info : http://www.activedir.org/List.aspx > List FAQ : http://www.activedir.org/ListFAQ.aspx > List archive: > http://www.mail-archive.com/activedir%40mail.activedir.org/ > List info : http://www.activedir.org/List.aspx > List FAQ : http://www.activedir.org/ListFAQ.aspx > List archive: http://www.mail-archive.com/activedir%40mail.activedir.org/ > > > > List info : http://www.activedir.org/List.aspx > List FAQ : http://www.activedir.org/ListFAQ.aspx > List archive: http://www.mail-archive.com/activedir%40mail.activedir.org/ > > > List info : http://www.activedir.org/List.aspx > List FAQ : http://www.activedir.org/ListFAQ.aspx > List archive: http://www.mail-archive.com/activedir%40mail.activedir.org/ > > > List info : http://www.activedir.org/List.aspx > List FAQ : http://www.activedir.org/ListFAQ.aspx > List archive: http://www.mail-archive.com/activedir%40mail.activedir.org/ > List info : http://www.activedir.org/List.aspx List FAQ : http://www.activedir.org/ListFAQ.aspx List archive: http://www.mail-archive.com/activedir%40mail.activedir.org/ List info : http://www.activedir.org/List.aspx List FAQ : http://www.activedir.org/ListFAQ.aspx List archive: http://www.mail-archive.com/activedir%40mail.activedir.org/
