11=======================================================
2- Frequently Asked Questions (FAQ) for PostgreSQL 7.1
2+ Frequently Asked Questions (FAQ) for PostgreSQL 7.3
33HP-UX Specific
44TO BE READ IN CONJUNCTION WITH THE NORMAL FAQ
55=======================================================
6- last updated: $Date: 2002/10/22 13 :46:59 $
6+ last updated: $Date: 2002/11/02 00 :46:23 $
77
88current maintainer: Tom Lane (tgl@sss.pgh.pa.us)
99original author: Tom Lane (tgl@sss.pgh.pa.us)
@@ -12,10 +12,7 @@ original author: Tom Lane (tgl@sss.pgh.pa.us)
1212Questions covered here:
13131.1)What do I need to install PostgreSQL on HP-UX?
14141.2)Anything special about the build/install procedure?
15- 1.3)yacc dies trying to process src/backend/parser/gram.y.
16- 1.4)Linking the main postgres executable fails, complaining that
17- there's no "alloca" function.
18- 1.5)OK, it seemed to build and install, but the regression test fails.
15+ 1.3)OK, it seemed to build and install, but the regression test fails.
1916
2017
2118----------------------------------------------------------------------
@@ -24,30 +21,27 @@ Section 1: Installing PostgreSQL
2421
25221.1)What do I need to install PostgreSQL on HP-UX?
2623
27- PostgreSQL 7.1 should work on Series 700/800 machines running HPUX 9.*,
28- 10.*, or 11.*, given appropriate system patch levels and build tools.
29- At least one developer routinely tests on HPUX 10.20, and we have reports
30- of successful install on HPUX 11. Releases before 7.1 were tested on
31- HPUX 9.03 and 9.05; 7.1 should still work there although no recent tests
32- have been reported.
24+ PostgreSQL 7.3 should work on Series 700/800 PA-RISC machines running
25+ HP-UX 10.X or 11.X, given appropriate system patch levels and build
26+ tools. At least one developer routinely tests on HPUX 10.20 and we
27+ have reports of successful installations on HPUX 11.00 and 11.11.
3328
3429Aside from the PostgreSQL source distribution, you will need GNU make
3530(HP's make will not do), and either GNU gcc or HP's full ANSI C compiler.
3631If you intend to build from CVS sources rather than a distribution tarball,
3732you will also need flex (GNU lex) and bison (GNU yacc).
3833
39- I'd also recommend making sure you are fairly up-to-date on HP patches,
40- particularly if you are using HPUX 9. At a minimum, if you are on HPUX 9,
41- you *must* have PHSS_4630 (libm update) or a successor patch; otherwise
42- Postgres' date/time functions will misbehave. On general principles you
43- should be current on libc and ld/dld patches, as well as compiler patches
44- if you are using HP's C compiler. (The only other presently known failure
45- from out-of-date system libraries is that on HPUX 10.10, the backend will
46- crash after the second error message in a session unless you have upgraded
47- libc to PHCO_16722 or later.)
34+ I'd also recommend making sure you are fairly up-to-date on HP
35+ patches. At a minimum, if you are building 64 bit binaries on on HPUX
36+ 11.11 you may need PHSS_26263 (11.11) or a successor patch otherwise
37+ initdb may hang:
4838
49- See HP's support websites, such as http://us-support.external.hp.com/,
50- for free copies of their latest patches.
39+ PHSS_26263 s700_800 11.11 ld(1) and linker tools cumulative patch
40+
41+ On general principles you should be current on libc and ld/dld
42+ patches, as well as compiler patches if you are using HP's C compiler.
43+ See HP's support sites such as http://itrc.hp.com and
44+ ftp://us-ffs.external.hp.com/ for free copies of their latest patches.
5145
5246
53471.2)Anything special about the build/install procedure?
@@ -64,51 +58,29 @@ The default install target location is /usr/local/pgsql, which
6458you might want to change to something under /opt. If so, use
6559the --prefix switch to configure.
6660
67-
68- 1.3)yacc dies trying to process src/backend/parser/gram.y.
69-
70- HP's yacc doesn't create its tables large enough to handle the Postgres
71- grammar (a lot of other vendors' yaccs have this problem too). The
72- preferred solution is to use GNU bison instead. If you don't want to
73- do that for some reason, it's possible to increase yacc's table sizes
74- enough to cope. With a pre-6.4 PostgreSQL grammar, I was able to get
75- HPUX 9's yacc to work by setting YFLAGS to
76- -d -Np2000 -Ns3000 -Nm100000 -Nl2000 -Na30000 -Nc10000
77- (You can edit YFLAGS either in the template file before running
78- configure, or in src/Makefile.global afterwards.) Later PostgreSQL
79- releases might require even larger tables, but this should do for
80- a starting point.
81-
82- Note that this shouldn't affect you if you are using a distribution
83- tarball, but it does matter if you pull the sources from the CVS server.
84- The CVS server does not store prebuilt yacc output files.
61+ If you are building on a PA-RISC 2.0 machine and want the compiled
62+ binaries to run on PA-RISC 1.1 machines you will need to specify
63+ +DAportable in CFLAGS.
8564
8665
87- 1.4)Linking the main postgres executable fails, complaining that
88- there's no "alloca" function.
89-
90- You're using an old version of GNU bison. Update to 1.50 or later,
91- and re-make the bison output files. Or build with gcc, or update to
92- HPUX 10, either of which will provide support for alloca.
93-
94-
95- 1.5)OK, it seemed to build and install, but the regression test fails.
66+ 1.3)OK, it seemed to build and install, but the regression test fails.
9667
9768There are several "expected failures" due to differences between HPUX
98- and the regression test reference platform used by the PostgreSQL group.
99- All of these should be compensated for by the regression test comparison
100- mechanism, with the possible exception of some low-order-digit differences
101- in the geometry tests (depending on whichcompiler and math library
102- versions you use) .
69+ and the regression test reference platform used by the PostgreSQL
70+ group. All of these should be compensated for by the regression test
71+ comparison mechanism, with the possible exception of some
72+ low-order-digit differences in the geometry tests, whichvary depending
73+ on which compiler and math library versions you use.
10374
104- Any other error is cause for suspicion. In particular, if you see
105- failures in the datetime test on HPUX 9, you probably forgot to
106- install the libm patch PHSS_4630 --- see item 1.1 above.
75+ Any other error is cause for suspicion.
10776
10877The parallel regression test script (gmake check) is known to lock up
109- when run under HP'sdefault Bourneshell, at least in HPUX 10.20. This
110- appears to be a shell bug, not thefault of the script. If you see that
111- the tests have stopped making progress and only a shell process is
112- consuming CPU, kill the shell and start over with
78+ when run under HP's Bourneshells: /usr/bin/sh and /sbin/sh. This is a
79+ known defect JAGad84609, thefix for which is not yet in any released
80+ HP-UX version or shell patches. To work around it, use ksh to run the
81+ regression script:
11382gmake SHELL=/bin/ksh check
114- to use ksh instead.
83+
84+ If you see that the tests have stopped making progress and only a shell
85+ process is consuming CPU, kill the shell process and start over with the
86+ above command.