June 25, 2014 and the recent massive system crash at the USPTO

The massive system crash at the USPTO began about August 15, 2018 and things were sort of almost back to normal around August 21.  During the recovery efforts USPTO posted updates that indicated that USPTO was in the position of having to reconstruct the Palm database, apparently by cobbling together various incremental backups to arrive at a reconstruction of what the database might have been like on some particular date and then applying more recent changes, eventually working toward what the contents of the Palm database would have been in the absence of whatever bad thing happened on the 15th.

Not that USPTO ever quite came clean on what exactly the bad thing was that happened to the Palm database on the 15th.  Maybe there was a ghost in the machine?

But anyway there were odd things that USPTO did around August 21, 2018 that related to a particular date from about four years earlier — June 25, 2014.  Members of the PAIR listserv noticed odd things relating to OCNs (Outgoing Correspondence Notifications).

It is recalled that when you set up a Customer Number in PAIR, you can pick whether your outgoing correspondence is sent on paper or whether USPTO will instead use OCNs instead of mailing out your correspondence on paper.  The way that OCNs work is that when there is correspondence that the USPTO wants you to know about, USPTO will send you an email (an OCN) and you will then click around in IFW (image file wrapper) to see the correspondence and print it out for yourself from PAIR.

And then if USPTO notices that you never actually got around to clicking on the document in IFW, USPTO will mail out a post card.  The point of the post card is to remind you to go to PAIR and open up IFW and click on the particular item of correspondence.

So anyway on August 21, 2018 the members of the PAIR listserv (which you should join, if you have not already done so) were discussing the most recent status of the PAIR system in the USPTO’s slow crawl toward normalcy after the massive crash.  And what several members of the listserv (including alert listserv member Rick Neifeld) noticed was that USPTO had sent out OCNs on August 21, 2018 relating to things that had happened on June 25, 2014.  Rick rather poetically referred to this OCN as a “ghost in the machine”.

In our case we received an email from USPTO on August 21, 2018 suggesting that we go to IFW and click to see a Notice of Allowance that had happened on June 25, 2014.  We did not actually go to IFW on August 21, 2018 to click on the 2014 document.  From what Rick posted to the PAIR listserv on August 21, 2018 I am guessing that likewise he did not actually go to IFW on August 21, 2018 to click on his 2014 document.

Anyway here is the fun part (fun?).  We have now on August 28, 2018 received a physical post card from the USPTO as part of the OCN workflow, reminding us that we need to go to IFW to click on our document that dates from June 25, 2014.  Yes, a ghost post card arrived in August of 2018 to tell us that we had better click on IFW to see something that happened in 2014.

These are post cards that USPTO mails by sending data files to USPS.  The USPS prints the post cards and mails them and sends a bill to USPTO.  So what’s going on here?  Did USPS print this post card and it got lost in the mail for four years and now the USPS is just getting around to delivering the post card that USPS printed four years ago?  No, I don’t think that’s it, given that the post card matches a corresponding spurious OCN email from the USPTO that USPTO emailed in 2018.

Yes, I imagine that part of what was going on when USPTO was reconstructing the Palm database as a cobbling-together of old and new incremental backups, there was some hiccup involving June 25, 2014.  I am guessing that whatever the bad thing was that happened on August 15, 2018 (a bad thing that even now USPTO has not quite come clean about), it led to USPTO going through older and older full backups of the Palm database until it found a full backup that it thought could be trusted.  And then applied the various incremental backups to that full backup to eventually arrive at a reconstruction of what the Palm database was when the (unnamed) bad thing happened on August 15, 2018.  And then this reconstructed Palm database was placed into production service on August 21, 2018.

And then there was an unintended consequence, namely a consequence of the OCN daily workflow that runs every day.  The OCN workflow happened.  The OCN workflow did what it always does once every day, which is to ask itself “what are the pieces of outgoing correspondence for which we emailed an OCN and we don’t see that the customer ever got around to clicking in IFW to look at it?”  And then the OCN workflow says “now that I have identified the things that never got clicked on in IFW, let’s send a data file to the USPS so that physical post cards will get mailed to those customers who never clicked on their stuff in IFW.”

And in this case, around August 24, 2018 the OCN workflow noticed that we did not go to IFW to click on the June 25, 2014 document.  (And, I would guess, noticed that Rick never went to IFW to click on his June 25, 2014 document.)  And then the OCN workflow sent a data file to the USPS on around August 24, 2018 asking the USPS to please mail out post cards.

Which then led to our receiving this physical post card in 2018, reminding us to go to IFW to click on the document from June 25, 2014.  (And, I would guess, has led to a similar physical post card getting mailed to Rick.)

A ghost post card.  A post card that got mailed in 2018 because USPTO’s reconstruction of the Palm database involved going back to June 25, 2014 (if my guess is right) for a full backup and then working forward with incremental backups to reconstruct the situation as of August of 2018.

Did you recently receive an OCN post card from USPTO in 2018, reminding you of an event that happened on June 25, 2014?  If so, please post a comment below.

6 Replies to “June 25, 2014 and the recent massive system crash at the USPTO”

  1. Are you seriously suggesting that the USPTO has not done a full backup of its databases in over four years (if ever)? And IIRC one of the official announcements mentioned that several TERAbytes of data had to be processed which, even shared over 10,000,000 patents, is a lot of data on each case.

    “How the human race ever survived…”.

Leave a Reply

Your email address will not be published. Required fields are marked *