> Hi,
>
> I have a CouchDB instance v1.3.0 with about 6 databases and one of them is
> used for historical
> reporting usually with 15 million records and about 30 GB to 45 GB disk space.
> This Sunday it just crashed first rejecting http requests and then even
> killing the couchDB
> process.
>
> It is running in a Windows Server 2008 with about 32 GB of ram and 8 CPU
> cores.
>
> While deleting a bunch of old records from the historical reports database
> this error
> message pops-up
>
> [Sun, 07 Dec 2014 18:41:43 GMT] [info] [<0.11190.28>] 15.128.50.71 - - DELETE
> /outbound-record-history/6ebaaebd8f0bcffd18121a227a3654f0?rev=1-bf5e3f4fb9cc341b2c7be7599a4f7a8f
>
> 200
> [Sun, 07 Dec 2014 18:41:43 GMT] [info] [<0.10928.28>] 15.128.50.71 - - DELETE
> /outbound-record-history/6ebaaebd8f0bcffd18121a227a3651d8?rev=1-7aca019d84960522a12a8d2a7c224e90
>
> 200
> [Sun, 07 Dec 2014 18:41:43 GMT] [info] [<0.13583.28>] 15.128.50.71 - - DELETE
> /outbound-record-history/6ebaaebd8f0bcffd18121a227a3655d2?rev=1-467254a30672ad3c56d59ada50e81789
>
> 200
> [Sun, 07 Dec 2014 18:41:43 GMT] [info] [<0.195.0>] Index shutdown by monitor
> notice
> for db: outbound-record-history idx: _design/RecordHistory
> [Sun, 07 Dec 2014 18:41:43 GMT] [info] [<0.202.0>] Index shutdown by monitor
> notice
> for db: outbound-record-history idx: _design/FileFormatter
> [Sun, 07 Dec 2014 18:41:43 GMT] [error] [<0.13583.28>] Uncaught error in HTTP
> request:
> {exit,
> {{badmatch,
> {error,eio}},
> [{couch_btree,
> '-write_node/3-lc$^0/1-0-‘,
[snip]
> Alejandro Santiago Femat
> Senior Consultant - Technical Solution Group - North LATAM
> Telefono +52.55.1101-6700 C +521.55.5505-1705
> [email protected]
> www.genesyslab.com
>
Hi Alejandro,
You’re not subscribed to the list, I moderated your message through (I hope).
https://github.com/erlang/otp/blob/maint/erts/emulator/drivers/win32/win_efile.c#L266-L268
I’m not a C expert, but reading the above this could occur on a network socket,
or as a
catch-all clause for “other” disk/storage related issues.
Given where this occurred, during b-tree updating, I’d assume you have a
failing disk, or
other storage-related problems. You should check event logs for anything
unusual, and
do a full compaction of your restored DB on another box to ensure that your
backup at least
was done prior to any corruption.
Do let us know what you find as the result of investigations.
A+, Dave
— sent from my Couch
