Re: Query about 'initdb' error - Mailing list pgsql-hackers

From Tom Lane
Subject Re: Query about 'initdb' error
Date
Msg-id 3716905.1616075742@sss.pgh.pa.us
Whole thread Raw
In response to Query about 'initdb' error  ("杨逸存" <1057206466@qq.com>)
List pgsql-hackers
"=?gb18030?B?0e7S3bTm?=" <1057206466@qq.com> writes:
>     When I do 'make' and 'make install', there is no warning or error infomation appeared. Compilation step
seemsto have no problem. Also, I use gdb to backtrace initdb, the result is as below: 

> performing post-bootstrap initialization ... Segmentation fault (core dumped)
> Program received signal SIGPIPE, Broken pipe.

> I am not able to fully recognize the insight of backtrace, but it's
> strange that it receives 'signal SIGPIPE, Broken pipe'.   
> Looking forward to your reply. Yours sincerely.

At this phase, initdb is just shoving SQL commands down a pipe to a
"standalone backend" that's doing the real work.  Evidently your
backend dumped core, and you need to be looking at that dump not
initdb itself.  Depending on how your machine is set up, you might
need to use initdb's --noclean option to keep it from throwing away
the incomplete data directory, as the backend might have dropped core
in there.

            regards, tom lane



pgsql-hackers by date:

Previous
From: Bruce Momjian
Date:
Subject: Re: fdatasync performance problem with large number of DB files
Next
From: Ibrar Ahmed
Date:
Subject: Re: [HACKERS] make async slave to wait for lsn to be replayed