[mythtv-users] Mythbackend suddenly deleting everything marked for deletion including unwatched

Mark Perkins perkins1724 at hotmail.com
Sat Feb 1 06:13:57 UTC 2014



> Date: Sat, 1 Feb 2014 05:19:05 +0100
> From: hikavdh at gmail.com
> To: mythtv-users at mythtv.org
> Subject: Re: [mythtv-users] Mythbackend suddenly deleting everything marked	for deletion including unwatched
> 
> Hoi Stephen,
> 
> Saturday, February 1, 2014, 4:58:43 AM, you wrote:
> 
> 
> > On 1/31/14, 10:52 PM, Hika van den Hoven wrote:
> >> Hoi Mark,
> >>
> >> Saturday, February 1, 2014, 4:39:58 AM, you wrote:
> >>
> >>
> >>
> >>>> On 1 Feb 2014, at 1:58 pm, "Hika van den Hoven" <hikavdh at gmail.com> wrote:
> >>>>
> >>>> Hoi R.G.,
> >>>>
> >>>> Saturday, February 1, 2014, 4:15:06 AM, you wrote:
> >>>>
> >>>>> Were the actual files deleted or only the entries in the database? These
> >>>>> two manifestations point to entirely different problems.
> >>>>> If the first, I would suspect the hard drive or controller, possibly caused
> >>>>> by a power surge. Thge second could also be caused by a power surge, and
> >>>>> possibly could be fixed by optimize-db.
> >>>>> But I am guessing here.
> >>>>> Geoff
> >>>>> Psvthis *should* come out as bottom boosted, if the setting does what it
> >>>>> sasys. New toy, my apologies if wrong.
> >>>>> G
> >>>>> Sent with AquaMail for Android
> >>>>> http://www.aqua-mail.com
> >>>>
> >>>>> On January 31, 2014 10:03:05 PM Hika van den Hoven <hikavdh at gmail.com> wrote:
> >>>>>> Hoi Discussion,
> >>>>>>
> >>>>>>  Some two hours ago I had the weirdest thing.
> >>>>>>  I was watching a recording and when it finished it sprang as usual
> >>>>>>  back to the list (unwatched recordings) and I saw some recordings
> >>>>>>  marked as unfindable. If had that before and restarting the viewer
> >>>>>>  solved it. But not this time. All recording not marked to keep where
> >>>>>>  gone, including a few unwatched and the one I had just watched.
> >>>>>>  When I realised what had happend I took the backend and the volume
> >>>>>>  offline. I'm now running an undelete with ext4magic on everything
> >>>>>>  recorded the last month, which seems to go fine. I only had first to
> >>>>>>  create space to store the files. The deleted files are some 180 Gb
> >>>>>>  and what I,m now undeleting is something in between a quarter and a
> >>>>>>  half of that, but I probably only want to restore the unwatched
> >>>>>>  ones.
> >>>>>>
> >>>>>>  Anybody an idea how this happened and how to restore the recordings
> >>>>>>  to the database. I could restore the database to last night, but
> >>>>>>  that leaves today out.
> >>>>>>
> >>>>>> --
> >>>>>> Tot Mails,
> >>>>>> Hika                          mailto:hikavdh at gmail.com
> >>>>>> Mijn Mailadres is gewijzigd. Zie hierboven!
> >>>>>>
> >>>>>> "Zonder hoop kun je niet leven
> >>>>>> Zonder leven is er geen hoop
> >>>>>> Het eeuwige dilemma
> >>>>>> Zeker als je hoop moet vernietigen om te kunnen overleven!"
> >>>>>>
> >>>>>> De lerende Mens
> >>>>>>
> >>>>>> _______________________________________________
> >>>>>> mythtv-users mailing list
> >>>>>> mythtv-users at mythtv.org
> >>>>>> http://www.mythtv.org/mailman/listinfo/mythtv-users
> >>>>
> >>>>> _______________________________________________
> >>>>> mythtv-users mailing list
> >>>>> mythtv-users at mythtv.org
> >>>>> http://www.mythtv.org/mailman/listinfo/mythtv-users
> >>>>
> >>>> They where physically deleted. To me it seems like it suddenly decided
> >>>> it needed much more space. To me it sounds like it got some wrong
> >>>> input about either needed space or available space.  Luckily it
> >>>> stopped at recordings marked to keep. The deleted recordings I hadn't
> >>>> yet watched are maybe 3, 5 at the most.
> >>>> I will look into the logs, but haven't had time. I first had to
> >>>> secure the volume and undelete what I can.
> >>>> I'm thinking of restoring last nights backup in parallel and copying
> >>>> the relevant records.
> >>>>
> >>> +1 to seeing what hints are in the logs. Would hate to restore the
> >>> backup from 24hrs ago for it to just do exactly the same thing 24hrs later.
> >>> I would also secure any recent backups and run a new backup ASAP
> >>> just to try and keep options available, don't want your best backup
> >>> to get overwritten by a junk one if your system is set to only keep a couple.
> >>> _______________________________________________
> >>> mythtv-users mailing list
> >>> mythtv-users at mythtv.org
> >>> http://www.mythtv.org/mailman/listinfo/mythtv-users
> >>
> >> My backups and the backupscript are on the data volume. What is the
> >> naming of the backupfiles? Among the deleted files there are 18 files
> >> like "mythconverge_backup-####.pl". where #### is a sequential number.
> >> I'll undelete them to be sure and I'll move the backupfiles for the
> >> future elsewhere with the ldap, tdb etc. backups.
> >> For now the undelete is just past halfway and when that finishes I go
> >> to bed! I was about to, when it happened.
> >>
> >> Tot mails,
> >>   Hika                            mailto:hikavdh at gmail.com
> >>
> >> "Zonder hoop kun je niet leven
> >> Zonder leven is er geen hoop
> >> Het eeuwige dilemma
> >> Zeker als je hoop moet vernietigen om te kunnen overleven!"
> >>
> >> De lerende Mens
> >> --
> >>
> >> _______________________________________________
> >>
> > I'll plead guilty to doing the same, but I have scripts that copy off
> > said data to a different storage unit.
> > Just in case the volume dies, then I still have a backup.
> 
> > As I recall, backups are named by date/time group. Hence, the latest
> > date/time on a 24 hour scale is the latest.
> 
> > I'll not go into my own undelete history, we'll suffice it to learn how
> > it works out for you.
> > My previous, unintentionally configured version didn't fare so well.
> > Later versions fared far better, by design and understanding of the
> > backup method.
> 
> > _______________________________________________
> > mythtv-users mailing list
> > mythtv-users at mythtv.org
> > http://www.mythtv.org/mailman/listinfo/mythtv-users
> 
> 
> Ok the log part when it went wrong:
> 
> 2014-02-01 00:21:06.659851 N [4582/4582] CoreContext autoexpire.cpp:264 (CalcParams) - AutoExpire: CalcParams(): Max required Free Space: 6.0 GB w/freq: 15 min
> 2014-02-01 00:21:06.932926 N [4582/4582] CoreContext autoexpire.cpp:264 (CalcParams) - AutoExpire: CalcParams(): Max required Free Space: 6.0 GB w/freq: 15 min
> 2014-02-01 00:22:18.035699 I [4582/18519] ProcessRequest mainserver.cpp:1395 (HandleAnnounce) - MainServer::ANN Monitor
> 2014-02-01 00:22:18.035713 I [4582/18519] ProcessRequest mainserver.cpp:1397 (HandleAnnounce) - adding: pc-Myth as a client (events: 0)
> 2014-02-01 00:22:18.037227 I [4582/18519] ProcessRequest mainserver.cpp:1395 (HandleAnnounce) - MainServer::ANN Monitor
> 2014-02-01 00:22:18.037239 I [4582/18519] ProcessRequest mainserver.cpp:1397 (HandleAnnounce) - adding: pc-Myth as a client (events: 1)
> 2014-02-01 00:22:36.937312 N [4582/4844] Expire autoexpire.cpp:641 (SendDeleteMessages) - Expiring 0 MB for 1 at 2014-01-31T23:21:06Z => "Pauw & Witteman"
> 2014-02-01 00:22:36.937383 N [4582/4844] Expire autoexpire.cpp:641 (SendDeleteMessages) - Expiring 9 MB for 1 at 2014-01-31T23:21:07Z => "Pauw & Witteman"
> 2014-02-01 00:23:21.703251 I [4582/4843] HouseKeeping housekeeper.cpp:221 (RunHouseKeeping) - Running housekeeping thread
> 2014-02-01 00:24:30.760948 E [4582/9611] Socket mainserver.cpp:5740 (connectionClosed) - Slave backend: pc-Myth no longer connected
> 2014-02-01 00:24:30.762352 I [4582/4841] Scheduler scheduler.cpp:2129 (HandleReschedule) - Reschedule requested for PLACE SlaveDisconnected
> 2014-02-01 00:24:30.783102 I [4582/4841] Scheduler mythdbcon.cpp:409 (PurgeIdleConnections) - New DB connection, total: 13
> 2014-02-01 00:24:30.809337 I [4582/4841] Scheduler scheduler.cpp:2242 (HandleReschedule) - Scheduled 22 items in 0.0 = 0.00 match + 0.00 check + 0.02 place
> 2014-02-01 00:27:02.482963 I [4582/18519] ProcessRequest mainserver.cpp:1395 (HandleAnnounce) - MainServer::ANN Monitor
> 2014-02-01 00:27:02.482976 I [4582/18519] ProcessRequest mainserver.cpp:1397 (HandleAnnounce) - adding: tzcheck as a client (events: 0)
> 2014-02-01 00:27:04.032495 I [4582/18519] ProcessRequest mainserver.cpp:1445 (HandleAnnounce) - adding: pc-Myth as a slave backend server
> 2014-02-01 00:27:04.033184 I [4582/4841] Scheduler scheduler.cpp:2129 (HandleReschedule) - Reschedule requested for PLACE SlaveConnected
> 2014-02-01 00:27:04.069020 I [4582/4841] Scheduler scheduler.cpp:2242 (HandleReschedule) - Scheduled 22 items in 0.0 = 0.00 match + 0.00 check + 0.01 place
> 2014-02-01 00:27:10.048127 N [4582/7296] Update autoexpire.cpp:264 (CalcParams) - AutoExpire: CalcParams(): Max required Free Space: 34359738372.0 GB w/freq: 3 min
> 2014-02-01 00:28:26.707281 I [4582/4843] HouseKeeping housekeeper.cpp:221 (RunHouseKeeping) - Running housekeeping thread
> 2014-02-01 00:32:36.057366 N [4582/4844] Expire autoexpire.cpp:264 (CalcParams) - AutoExpire: CalcParams(): Max required Free Space: 34359738372.0 GB w/freq: 3 min
> 2014-02-01 00:32:36.520675 N [4582/4844] Expire autoexpire.cpp:641 (SendDeleteMessages) - Expiring 1942 MB for 1 at 2013-12-11T20:25:00Z => "Case Sensitive"
> 2014-02-01 00:32:36.520721 N [4582/4844] Expire autoexpire.cpp:641 (SendDeleteMessages) - Expiring 1310 MB for 4 at 2013-12-12T13:35:00Z => "New tricks"
> 2014-02-01 00:32:36.520752 N [4582/4844] Expire autoexpire.cpp:641 (SendDeleteMessages) - Expiring 2025 MB for 2 at 2013-12-12T20:10:00Z => Zembla
> 
> As you can see the 'Max required Free Space' suddenly changed from 6.0
> Gb as I had set it to 34359738372.0 GB which is astronomical and
> probably the max allowed value in the database. Why?...
> It took about an hour to delete all it's allowed and with this setting
> it will keep deleting everything set to autoexpire.
> I'm now into mythtv-setup to try to correct it
> 

Yes that's what has done it. But what caused it to suddenly change value?

The slave backend disconnected / reconnected just before it all went down. As i see it, the slave disconnected, there was a quick reschedule as a result. Then the slave reconnected 2.5 minutes later with another quick reschedule as the slave was available again. But then 6 seconds later Update Autoexpire ran which is what returned the dodgy free space requirement. What caused Update Autoexpire to run? I checked 3 days of my logs and it's not run once (but I have no slave BE either). Maybe the slave disconnecting / reconnecting tripped up the Update Autoexpire call? Hopefully someone who knows how these bits all tie together will comment.

Is pc-Myth a combined FE/Slave-BE?
 		 	   		  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.mythtv.org/pipermail/mythtv-users/attachments/20140201/6fc4135a/attachment.html>


More information about the mythtv-users mailing list