agora inbox for postgres@postgres.berkeley.edu  
help / color / mirror / Atom feed
(no subject)
6+ messages / 5 participants
[nested] [flat]

* (no subject)
@ 1996-03-05 13:33  Avi Cohen Stuart <avi@baan.nl>
  0 siblings, 3 replies; 6+ messages in thread

From: Avi Cohen Stuart @ 1996-03-05 13:33 UTC (permalink / raw)
  To: legacy

To all,

I've succeeded to compile the complete postgres v4.2 for my HP9000 425t
under HPUX 9.10 but I have some problems during the initdb phase:

I'm getting the following messages:
fmgr_info: function 0: cache lookup failed.

12 like these in a row. I've tracked down the guilty lines responsable
for these errors:
they occur in the 'declare index on ....' at the end of the 
local1_template1.bki (or something) file.

It appears that the SearchSysCacheTuple() function fails.

Is there still a postgres expert around who can tell me where I should
look? 

Avi.

==============================================================================
   To add/remove yourself to/from the POSTGRES mailing list: send mail with 
   the subject line ADD or DEL to "postgres-request@postgres.Berkeley.EDU".
   If this fails, send mail to "post_questions@postgres.Berkeley.EDU" and
   a human will deal with it.  DO NOT post to the "postgres" mailing list.
==============================================================================
              URL: http://s2k-ftp.CS.Berkeley.EDU:8000/postgres/



^ permalink  raw  reply  [nested|flat] 6+ messages in thread

* Re: your mail
@ 1996-03-05 14:50  Avi Cohen Stuart <avi@baan.nl>
  parent: Thomas Krebs <krebs@faps.uni-erlangen.de>
  0 siblings, 0 replies; 6+ messages in thread

From: Avi Cohen Stuart @ 1996-03-05 14:50 UTC (permalink / raw)
  To: krebs@faps.uni-erlangen.de; +Cc: legacy

   > 
   > I'm getting the following messages:
   > fmgr_info: function 0: cache lookup failed.
   > 
   > 12 like these in a row. I've tracked down the guilty lines responsable
   > for these errors:
   > they occur in the 'declare index on ....' at the end of the 
   > local1_template1.bki (or something) file.
   > 
   > It appears that the SearchSysCacheTuple() function fails.
   > 
   > Is there still a postgres expert around who can tell me where I should
   > look? 
   > 
   Some time ago I tried to port Postgres 4.2 to Domain/OS and I encountered
   the same problem. I tracked it down so far that there is a problem when
   reading a block from the disk.
   I must admit that I have not been able to fix that problem... :-(

Do you still remember which function caused this problem?
I may try to crack the problem then. It would be a shame if I don't get
this baby running since it looks very powerful to me and I want to have
a database with an SQL like language.

Avi.



==============================================================================
   To add/remove yourself to/from the POSTGRES mailing list: send mail with 
   the subject line ADD or DEL to "postgres-request@postgres.Berkeley.EDU".
   If this fails, send mail to "post_questions@postgres.Berkeley.EDU" and
   a human will deal with it.  DO NOT post to the "postgres" mailing list.
==============================================================================
              URL: http://s2k-ftp.CS.Berkeley.EDU:8000/postgres/



^ permalink  raw  reply  [nested|flat] 6+ messages in thread

* Re: your mail
@ 1996-03-05 15:48  Thomas Krebs <krebs@faps.uni-erlangen.de>
  parent: Avi Cohen Stuart <avi@baan.nl>
  2 siblings, 1 reply; 6+ messages in thread

From: Thomas Krebs @ 1996-03-05 15:48 UTC (permalink / raw)
  To: avi@baan.nl; +Cc: legacy

> 
> To all,
> 
> I've succeeded to compile the complete postgres v4.2 for my HP9000 425t
> under HPUX 9.10 but I have some problems during the initdb phase:
> 
> I'm getting the following messages:
> fmgr_info: function 0: cache lookup failed.
> 
> 12 like these in a row. I've tracked down the guilty lines responsable
> for these errors:
> they occur in the 'declare index on ....' at the end of the 
> local1_template1.bki (or something) file.
> 
> It appears that the SearchSysCacheTuple() function fails.
> 
> Is there still a postgres expert around who can tell me where I should
> look? 
> 
Some time ago I tried to port Postgres 4.2 to Domain/OS and I encountered
the same problem. I tracked it down so far that there is a problem when
reading a block from the disk.
I must admit that I have not been able to fix that problem... :-(

Thomas
--
Thomas Krebs
Department for Manufacturing Automation and Production Systems FAPS
University of Erlangen
Egerlandstr. 7-9
91058 Erlangen
Tel.: +49 (0)9131/85-8740
Fax:  +49 (0)9131/302528
http://www.faps.uni-erlangen.de:1200/persons/krebs.html

==============================================================================
   To add/remove yourself to/from the POSTGRES mailing list: send mail with 
   the subject line ADD or DEL to "postgres-request@postgres.Berkeley.EDU".
   If this fails, send mail to "post_questions@postgres.Berkeley.EDU" and
   a human will deal with it.  DO NOT post to the "postgres" mailing list.
==============================================================================
              URL: http://s2k-ftp.CS.Berkeley.EDU:8000/postgres/



^ permalink  raw  reply  [nested|flat] 6+ messages in thread

* (no subject)
@ 1996-03-05 16:53  Jolly Chen <jolly@postgres.Berkeley.EDU>
  parent: Avi Cohen Stuart <avi@baan.nl>
  2 siblings, 0 replies; 6+ messages in thread

From: Jolly Chen @ 1996-03-05 16:53 UTC (permalink / raw)
  To: avi@baan.nl; +Cc: legacy



> I've succeeded to compile the complete postgres v4.2 for my HP9000 425t
> under HPUX 9.10 but I have some problems during the initdb phase:
> 

There is no support for postgres v4.2 any more.  Try the SQL-ized, improved
version of postgres, called postgres95.  The latest version is 1.01.
See our web page at
	http://s2k-ftp.cs.berkeley.edu:8000/postgres95
or ftp it from
	ftp://s2k-ftp.cs.berkeley.edu/pub/postgres95

> Is there still a postgres expert around who can tell me where I should
> look? 
> 

The "experts" are hanging out in the new mailing list
postgres95@postgres95.vnet.net.  (for more info, send mail to
majordomo@postgres95.vnet.net with 'info postgres95' in the body)

- Jolly

==============================================================================
   To add/remove yourself to/from the POSTGRES mailing list: send mail with 
   the subject line ADD or DEL to "postgres-request@postgres.Berkeley.EDU".
   If this fails, send mail to "post_questions@postgres.Berkeley.EDU" and
   a human will deal with it.  DO NOT post to the "postgres" mailing list.
==============================================================================
              URL: http://s2k-ftp.CS.Berkeley.EDU:8000/postgres/



^ permalink  raw  reply  [nested|flat] 6+ messages in thread

* (no subject)
@ 1996-03-05 21:06  V.Grabner <zen@bononunu.gams.co.at>
  parent: Avi Cohen Stuart <avi@baan.nl>
  2 siblings, 1 reply; 6+ messages in thread

From: V.Grabner @ 1996-03-05 21:06 UTC (permalink / raw)
  To: avi@baan.nl; +Cc: legacy

> To all,
> 
> I've succeeded to compile the complete postgres v4.2 for my HP9000 425t
> under HPUX 9.10 but I have some problems during the initdb phase:
> 
> I'm getting the following messages:
> fmgr_info: function 0: cache lookup failed.
> 
> 12 like these in a row. I've tracked down the guilty lines responsable
> for these errors:
> they occur in the 'declare index on ....' at the end of the 
> local1_template1.bki (or something) file.
> 
> It appears that the SearchSysCacheTuple() function fails.
> 
> Is there still a postgres expert around who can tell me where I should
> look? 

We had the same problem when using hash indices. Hash indices
alsways crashed the database when creating an index or inserting 
> 50 rows into the table.

Do not use them, use btree indices instead.

I tried to trace the problem but I did not find it. I found
out it always happend during the split of an hash page but I
never found out why. Without documentation I was lost.

--zen
 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
 Vinzenz Grabner                             Voice   : + 43 1 817 62 30-15
 G.A.M.S. GmbH.                              Fax     : + 43 1 817 62 30-17
 Schoenbrunnerstr. 179/II/4.St.
 A-1120 Vienna/Austria/Europe                e-mail  : zen@gams.co.at
 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -


==============================================================================
   To add/remove yourself to/from the POSTGRES mailing list: send mail with 
   the subject line ADD or DEL to "postgres-request@postgres.Berkeley.EDU".
   If this fails, send mail to "post_questions@postgres.Berkeley.EDU" and
   a human will deal with it.  DO NOT post to the "postgres" mailing list.
==============================================================================
              URL: http://s2k-ftp.CS.Berkeley.EDU:8000/postgres/



^ permalink  raw  reply  [nested|flat] 6+ messages in thread

* Re: hash indices
@ 1996-03-05 23:46  Paul M. Aoki <aoki@cs.berkeley.edu>
  parent: V.Grabner <zen@bononunu.gams.co.at>
  0 siblings, 0 replies; 6+ messages in thread

From: Paul M. Aoki @ 1996-03-05 23:46 UTC (permalink / raw)
  To: V.Grabner <zen@bononunu.gams.co.at>; +Cc: avi@baan.nl; legacy

"V.Grabner" <zen@bononunu.gams.co.at> writes:
> We had the same problem when using hash indices. Hash indices
> alsways crashed the database when creating an index or inserting 
> > 50 rows into the table.
> 
> Do not use them, use btree indices instead.
> 
> I tried to trace the problem but I did not find it. I found
> out it always happend during the split of an hash page but I
> never found out why. Without documentation I was lost.

the hash index split code is really, really broken in 4.2.  it doesn't
surface until you hit some fill factor (i.e., get a bunch of page
collisions), but after that the hash index is corrupt.  i think the
fixes (bug fixes, taking out the bogus "high concurrency" locking,
rewriting the split code) made it into some release of pg95 or
another..
--
  Paul M. Aoki          |  University of California at Berkeley
  aoki@CS.Berkeley.EDU  |  Dept. of EECS, Computer Science Division (#1776) 
                        |  Berkeley, CA 94720-1776

==============================================================================
   To add/remove yourself to/from the POSTGRES mailing list: send mail with 
   the subject line ADD or DEL to "postgres-request@postgres.Berkeley.EDU".
   If this fails, send mail to "post_questions@postgres.Berkeley.EDU" and
   a human will deal with it.  DO NOT post to the "postgres" mailing list.
==============================================================================
              URL: http://s2k-ftp.CS.Berkeley.EDU:8000/postgres/



^ permalink  raw  reply  [nested|flat] 6+ messages in thread


end of thread, other threads:[~1996-03-05 23:46 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz follow: Atom feed)
-- links below jump to the message on this page --
1996-03-05 13:33  Avi Cohen Stuart <avi@baan.nl>
1996-03-05 15:48 ` Re: your mail Thomas Krebs <krebs@faps.uni-erlangen.de>
1996-03-05 14:50   ` Re: your mail Avi Cohen Stuart <avi@baan.nl>
1996-03-05 16:53 ` Jolly Chen <jolly@postgres.Berkeley.EDU>
1996-03-05 21:06 ` V.Grabner <zen@bononunu.gams.co.at>
1996-03-05 23:46   ` Re: hash indices Paul M. Aoki <aoki@cs.berkeley.edu>

This inbox is served by agora; see mirroring instructions
for how to clone and mirror all data and code used for this inbox