couchdb-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Matthew Sinclair-Day <>
Subject Re: Couch crash report
Date Mon, 15 Nov 2010 19:03:33 GMT

A bit of a miscommunication on my end--the database was blown 
away.  We're in the process of trying to repeat the problem--a 
good amount of steady load, followed by inactivity allowing for 
the system to quiet down, then a single request.

I noticed this block in the log.  Not sure whether it was 
included in the attachment sent earlier:

[error] [<0.1111.10>] ** Generic server <0.1111.10> terminating
** Last message in was {pread_iolist,29142}
** When Server state == {file,{file_descriptor,prim_file,{#Port<0.50710>,24}},

I forgot to mention we are running on top of VMWare, in case 
that matters.


On 11/15/10 at 1:21 PM, (Robert Newson) wrote:

>Thanks! We've tracked it to;
>read_raw_iolist_int(#file{fd=Fd, tail_append_begin=TAB}, Pos, Len) ->
>BlockOffset = Pos rem ?SIZE_BLOCK,
>TotalBytes = calculate_total_read_len(BlockOffset, Len),
>{ok, <<RawBin:TotalBytes/binary>>} = file:pread(Fd, Pos, TotalBytes),
>if Pos >= TAB ->
>{remove_block_prefixes(BlockOffset, RawBin), Pos + TotalBytes};
>true ->
><<ReturnBin:Len/binary, _/binary>> = RawBin,
>{[ReturnBin], Pos + Len}
>The only reasonable cause is some issue with the file system or disk,
>specifically that file:pread returned less than TotalBytes number of
>bytes, causing the badmatch.
>Is the database in question confidential or private in any way? We
>might like to examine it in more detail if not.
>On Mon, Nov 15, 2010 at 5:46 PM, Matthew Sinclair-Day <> wrote:
>>On 11/15/10 at 12:39 PM, (Robert Newson) wrote:
>>> exception exit: {{badmatch,{ok
>>> <snip...>
>>> can you unsnip the rest of the badmatch exception text?
>>> B.
>>Unsnipped as an attachment (unless the mailing list strips it).

View raw message