Thread

1. Loosing files after backend crash

   PostgreSQL Bugs List <pgsql-bugs@postgresql.org> — 2001-04-02T21:45:28Z

   Konstantin Solodovnikov (ks@tcnet.ru) reports a bug with a severity of 2
   The lower the number the more severe it is.
   
   Short Description
   Loosing files after backend crash
   
   Long Description
   After a backend crash the file for my sequence cannot be found.
   
   I use PostgreSQL 7.1RC1 compiled by GCC 2.95.2 on FreeBSD 4.2-STABLE and that's what I encountered.
   
   During execution of some sql code backend crashed (the reason for it's crashing is not the actual topic of this report).
   I successfully re-run psql without restarting the postmaster.
   And tried to delete the sequence, which i created in the previous session.
   That's what pgsql told me:
   ---8<---
   Netflow_Test=# drop SEQUENCE seq_i___data_buffer ;
   ERROR:  cannot open seq_i___data_buffer: No such file or directory
   ---8<---
   I examined the system tables (ps_class) and found the tuple, describing my sequence.
   The "relfilenode" attribute of the tuple contained the value "326210". I tried to find the file with that name in the database directory - it really was not there.
   Thus, the backend crash has led not only to impossibility to complete my task, but also to loosing some data. The sequence was actually created in an already committed transaction, not in the one, that crashed.
   Another problem is that I cannot fix the situation (drop that sequence) by standard PGSQL means.
   I had to create an empty file with that name ("326210") for "DROP SEQUENCE" command to successfully complete.
   
   As i guess, the situation is not reprodusable. Backend crashed many times before w/o such problems.
   
   A little additional question/proposition: can there be any tool, which could find and resolve the errors in the system tables?
   
   Best Regards, 
   Konstantin Solodovnikov.
   
   Sample Code
   
   
   No file was uploaded with this report
   
   
   

2. Re: Loosing files after backend crash

   Tom Lane <tgl@sss.pgh.pa.us> — 2001-04-02T23:34:42Z

   pgsql-bugs@postgresql.org writes:
   > During execution of some sql code backend crashed (the reason for it's
   > crashing is not the actual topic of this report).
   > And tried to delete the sequence, which i created in the previous session.
   > That's what pgsql told me:
   > ---8<---
   > Netflow_Test=# drop SEQUENCE seq_i___data_buffer ;
   > ERROR:  cannot open seq_i___data_buffer: No such file or directory
   
   Was that a system-wide crash, or just a backend crash?  It seems
   to me that your kernel must have lost that file for you; Postgres
   wouldn't have gone out and deleted it.
   
   > Another problem is that I cannot fix the situation (drop that
   > sequence) by standard PGSQL means.  I had to create an empty file with
   > that name ("326210") for "DROP SEQUENCE" command to successfully
   > complete.
   
   This used to work pre-7.1, and it seems to me it should still work;
   relcache.c is being inappropriately paranoid about finding the file.
   I have corrected that part of the problem.
   
   			regards, tom lane