sks-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Sks-devel] unmergeable keys (2005-01-16)


From: Jason Harris
Subject: Re: [Sks-devel] unmergeable keys (2005-01-16)
Date: Sun, 16 Jan 2005 23:20:45 -0500
User-agent: Mutt/1.4.2.1i

On Sun, Jan 16, 2005 at 10:35:27PM -0500, Yaron Minsky wrote:
> On Sun, 16 Jan 2005 20:39:57 -0500, Jason Harris <address@hidden> wrote:
> > 
> > The SKS hashes below keep showing up in my diff-62.94.26.10_11371.txt.
> > I've disabled keyserver.linux.it as a sync. peer temporarily, until
> > these "permanent diffs" are resolved.
> > 
> > Importing the first and last hashes/keys into GPG and sending them to
> > my SKS server updated the keys in my db to have matching hashes, so
> > the keys themselves are probably fine.  I assume my PTree is also
> > good since I don't get such diffs with my other sync. peers.
> > 
> > I don't see how this is simply a case of PTree "corruption."  Yaron?
> 
> First of all, it sounds like the host in question's keyserver does
> indeed know about these keys, since you got the first and last key,
> presumably by asking that keyserver for the keys in question.  So no,
> it doesn't sound like the KDB and PTree are de-sync'd.  So, the

I was able to manually fetch all hashes from keyserver.linux.it using
op=hget URLs.

> question is, why can't the synchronization complete.  Perhaps you
> could post a snipper from the log showing a synchronizatoin attempt. 

It is like they weren't being fed to my "sks db" process, because I
don't see any evidence of them in db.log.  All I have are several
sessions like this in recon.log:

  2005-01-17 00:57:00 Beginning recon as server, client: <ADDR_INET 
62.94.26.10:57542>
  2005-01-17 00:57:00 Joining reconciliation
  2005-01-17 00:57:02 Reconciliation complete
  2005-01-17 00:57:02 113 hashes recovered from <ADDR_INET 62.94.26.10:11371>
  2005-01-17 00:57:02 Disabling gossip
  2005-01-17 00:57:18 Requesting 100 missing keys from <ADDR_INET 
62.94.26.10:11371>, starting with 005167E46F471CA63E40CD357E62A9A8
  2005-01-17 00:57:18 Error getting missing keys: Failure("Error parsing 
headers: no colon found")
  2005-01-17 00:57:19 Requesting 13 missing keys from <ADDR_INET 
62.94.26.10:11371>, starting with D6F36EF340D7AA4CAB32AC415FC86AB5
  2005-01-17 00:57:19 Error getting missing keys: Failure("Error parsing 
headers: no colon found")
  2005-01-17 00:57:20 Enabling gossip

> That might clarify why the synchronization is failing.  It sounds like
> keyserver.linux.it has some extra keys which for some reason aren't
> getting sent out, and it would be good to figure out why.

I fed the manually retrieved keys to my pks server, where they made their
way back into the SKS network sans photo IDs.  Now, most of the 111 hashes
are found on my SKS server.  Adding keyserver.linux.it back as a peer,
I'm down to 3 "permanent diffs" with it:

  %cat diff-62.94.26.10_11371.txt
  7910D41F23C7952B83AF2B0634CA988A
  7DD987395EF1A3DC85B98067BCBB0172
  B211769C2C0CEBDCD066881569BF85B5

recon.log says the keys are being received:

  2005-01-17 04:06:07 Hashes recovered from <ADDR_INET 62.94.26.10:11371>
  2005-01-17 04:06:07     74D6A858423FFFAA4D8435878EFF32A9
  2005-01-17 04:06:07     7910D41F23C7952B83AF2B0634CA988A
  2005-01-17 04:06:07     7DD987395EF1A3DC85B98067BCBB0172
  2005-01-17 04:06:07     B211769C2C0CEBDCD066881569BF85B5

but db.log only mentions the 1 mergeable key in the bunch:

  2005-01-17 04:07:12 0 potential merges found for keyid BCD6D5AD
  2005-01-17 04:07:12 1 updates found before filtering
  2005-01-17 04:07:12 Applying 1 changes
  2005-01-17 04:07:12 Adding hash 74D6A858423FFFAA4D8435878EFF32A9

-- 
Jason Harris           |  NIC:  JH329, PGP:  This _is_ PGP-signed, isn't it?
address@hidden _|_ web:  http://keyserver.kjsl.com/~jharris/
          Got photons?   (TM), (C) 2004

Attachment: pgp4bvnjZg9Ds.pgp
Description: PGP signature


reply via email to

[Prev in Thread] Current Thread [Next in Thread]