Hello there
I had some problems in the beginning of 2007 and I got some direction from
Agoston Horvath (thanks btw). It kinda solved some of the problems, but I
had to run the DB as a static DB to avoid crashes...
Anyway, due to some hardware problems we decided to rebuild the whois
server (version 3.3) on a new box. Agoston Horvath wrote that there would
be a new whois server avail. and 3.3 was old... but I can only find ver.
3.3.
Question: Is ver. 3.3 the latest avil. version? If not, where do I pickup
the latest and best (not via you CVS link ...) ?
version 3.3:
Compiling took ages due to a stripped RHEL (prerequisites not in place)
but it looks like I've succeeded at last.
Loading SAMPLE went OK - but I have problems with RIPE DB (would it be an
idea for future releases to store RIPE.CURRENTSERIAL within the ripe.db.gz
file ...just for the ease of it ?)
First - I'm running a MySQL 4.1.20 but that is the only *big* difference
in the list of prerequisites.
Second - a minor hack applied to <installdir>/bin/SQL/create.first-stage
since I apparently had som problems in the "auto" creation of the
RIPADMIN.
I added the following vars. to the script which helped (somehow these
vars. was not passed on from my rip.config file:
RIPADMIN_DBHOST=localhost
RIPADMIN_DBPORT=3306
RIPADMIN_DBUSER=<my_db_user>
RIPADMIN_DBPASS=<my_passwd>
RIPADMIN_DBNAME=RIPADMIN
Running make_db with option -0 runs fine :-)
Running make_db (in the ex. below with DEBUG enabled) with option -1 fails:
nohup /usr/local/whoisserver/bin/make_db -c
/usr/local/whoisserver/conf/rip.config -s RIPE -1 2>&1
[cut..cut]
[0] 0.03s OK [aut-num:AS8387][U:ADD]
[0] 0.04s OK [aut-num:AS1205][U:ADD]
[0] 0.02s OK [aut-num:AS15631][U:ADD]
[0] 0.02s OK [aut-num:AS8925][U:ADD]
[0] 0.06s OK [aut-num:AS12565][U:ADD]
[0] 0.04s OK [aut-num:AS8787][U:ADD]
[0] 0.03s OK [aut-num:AS15646][U:ADD]
[0] 0.02s OK [aut-num:AS15649][U:ADD]
[0] 0.02s OK [aut-num:AS15650][U:ADD]
[0] 0.02s OK [aut-num:AS15651][U:ADD]
[0] 0.03s OK [aut-num:AS8602][U:ADD]
[0] 0.04s OK [aut-num:AS13211][U:ADD]
[0] 0.04s OK [aut-num:AS12508][U:ADD]
[0] 0.03s OK [aut-num:AS13004][U:ADD]
[0] 0.06s OK [aut-num:AS15654][U:ADD]
[0] 0.04s OK [aut-num:AS15655][U:ADD]
[0] 0.03s OK [aut-num:AS15656][U:ADD]
[0] 0.03s OK [aut-num:AS15508][U:ADD]
*** ERROR: Loader failure [139]. Exiting\n
+ '[' 1 -ne 0 ']'
+ error_exit 'Error loading database=RIPE for source=RIPE,
/usr/local/whoisserver/bin/make_db exiting'
+ error Error loading database=RIPE for source=RIPE,
/usr/local/whoisserver/bin/make_db exiting
+ echo '************** ERROR ***************'
************** ERROR ***************
++ eval /bin/date ''\''+%H:%M:%S'\'''
+++ /bin/date +%H:%M:%S
+ echo '*** ' 12:31:36 Error loading database=RIPE for source=RIPE,
/usr/local/whoisserver/bin/make_db exiting
*** 12:31:36 Error loading database=RIPE for source=RIPE,
/usr/local/whoisserver/bin/make_db exiting
+ echo '************************************'
************************************
+ echo
+ exit 1
[end..end]
I've looked for the error code [139] in the src, but with no
*hey-here-is-the-solution* result.
Question: What does the error code imply ?
The exit/error messages on tty says:
[root@powder DB]# 0 rows affected [ INSERT mnt_routes SELECT 0,
19974, mntner.object_id, 2 FROM mntner WHERE
mntner.mntner='planetwebhost-mnt' ]
*** died: +1576 modules/ud/ud_core.c
*** Backtrace:
Using host libthread_db library "/lib/tls/libthread_db.so.1".
[Thread debugging using libthread_db enabled]
[New Thread -1208465184 (LWP 30785)]
0x005777a2 in _dl_sysinfo_int80 () from /lib/ld-linux.so.2
./load_all.sh: line 15: 30782 Broken pipe ${UNZIP1}
${object_file}
30783 Exit 141 | eval ${FILTER}
30785 Segmentation fault | $LOADER -L $1 -p ${PROPERTIES} -s
${SOURCE}
+ check_error 'Loader failure' 139
+ '[' 139 -ne 0 ']'
+ echo '*** ERROR: Loader failure [139]. Exiting\n'
+ exit 1
I thought there was an error/inconsistency in the input file (ripe.db.gz)
but ripe.db.gz from yesterday (May 14th) causes the same problem. So I
guess the problem lies somewhere in my binaries (the loader) or in mysql ?
Running make_db with the -2 option crashes as well - but it succeeds
loading much more data. Apparently it fails trying to update:
[root@powder DB]# tail -30 RIPE-load-type-2-run.log
[0] 0.03s OK [as-block:AS3416 - AS3459][U:UPD]
[0] 0.03s OK [as-block:AS3412 - AS3415][U:UPD]
[0] 0.03s OK [as-block:AS4383 - AS4404][U:UPD]
[0] 0.03s OK [as-block:AS4405 - AS4430][U:UPD]
[0] 0.03s OK [as-block:AS34816 - AS35839][U:UPD]
[0] 0.03s OK [as-block:AS35840 - AS36863][U:UPD]
[0] 0.03s OK [as-block:AS30720 - AS30979][U:UPD]
[0] 0.03s OK [as-block:AS30981 - AS30981][U:UPD]
[0] 0.03s OK [as-block:AS31000 - AS31743][U:UPD]
[0] 0.04s OK [as-block:AS30980 - AS30999][U:UPD]
[0] 0.03s OK [as-block:AS36864 - AS37887][U:UPD]
[0] 0.03s OK [as-block:AS37888 - AS38911][U:UPD]
[0] 0.03s OK [as-block:AS38912 - AS39935][U:UPD]
[0] 0.03s OK [as-block:AS39936 - AS40959][U:UPD]
[0] 0.03s OK [as-block:AS40960 - AS43007][U:UPD]
*** ERROR: Loader failure [139]. Exiting\n
+ '[' 1 -ne 0 ']'
+ error_exit 'Error loading database=RIPE for source=RIPE,
/usr/local/whoisserver/bin/make_db exiting'
+ error Error loading database=RIPE for source=RIPE,
/usr/local/whoisserver/bin/make_db exiting
+ echo '************** ERROR ***************'
************** ERROR ***************
++ eval /bin/date ''\''+%H:%M:%S'\'''
+++ /bin/date +%H:%M:%S
+ echo '*** ' 13:54:00 Error loading database=RIPE for source=RIPE,
/usr/local/whoisserver/bin/make_db exiting
*** 13:54:00 Error loading database=RIPE for source=RIPE,
/usr/local/whoisserver/bin/make_db exiting
+ echo '************************************'
************************************
+ echo
+ exit 1
The tty output for the option "-2" run says:
[cut..cut]
+ echo 'finished create.second-stage...'
+ '[' 2 -eq 2 ']'
++ date
+ echo Tue May 15 13:53:43 CEST 2007
+ echo 'Loading tables - 2 of 2 pass'
+ loadpass 2
+ for object_file in '${LIST}'
++ date
+ echo Tue May 15 13:53:43 CEST 2007 Loading
/usr/local/whoisserver/var/tmp/load/RIPE/ripe.db.gz
+ test -f /usr/local/whoisserver/var/tmp/load/RIPE/ripe.db.gz
+ check_error '/usr/local/whoisserver/var/tmp/load/RIPE/ripe.db.gz not
found' 0
+ '[' 0 -ne 0 ']'
+ gzip -cd /usr/local/whoisserver/var/tmp/load/RIPE/ripe.db.gz
+ eval cat
++ cat
+ /usr/local/whoisserver/bin/loader -L 2 -p
/usr/local/whoisserver/conf/rip.config -s RIPE
D: Making SQL connection to RIPE@localhost ...OK
starting processing stream
Object cannot be retrieved:[as-block: AS0 - AS0
descr: IANA held ASN block
remarks: These AS numbers are reserved by IANA
org: ORG-IANA1-RIPE
admin-c: IANA1-RIPE
tech-c: IANA1-RIPE
mnt-by: RIPE-DBM-MNT
mnt-lower: RIPE-DBM-MNT
changed: ripe-dbm(a)ripe.net 20070101
source: RIPE
]
*** died: +1144 modules/ud/ud_process_stream.c
*** Backtrace:
Using host libthread_db library "/lib/tls/libthread_db.so.1".
[Thread debugging using libthread_db enabled]
[New Thread -1208882976 (LWP 31113)]
0x005777a2 in _dl_sysinfo_int80 () from /lib/ld-linux.so.2
./load_all.sh: line 15: 31111 Broken pipe ${UNZIP1}
${object_file}
31112 Exit 141 | eval ${FILTER}
31113 Segmentation fault | $LOADER -L $1 -p ${PROPERTIES} -s
${SOURCE}
+ check_error 'Loader failure' 139
+ '[' 139 -ne 0 ']'
+ echo '*** ERROR: Loader failure [139]. Exiting\n'
+ exit 1
[end..end]
So...the loader errorcode is still "139" but it fails during UDP, not ADD
Any hints ? I'm a little lost ;-(
Should I give it a try with the "spilt" version of the current ripe.db.gz
? If yes, is the right way to point at the split files like this (in
./bin/make_db):
# List of object files (wildcards are allowed)
#LIST=${LIST:-$OBJDIR/sample.db.gz}
#LIST=${LIST:-$OBJDIR/ripe.db.gz}
LIST=${LIST:-$OBJDIR/*.db.gz}
I hope there is enough (and not to much...) info. to help pin-pointing
where to look.
Best regards,
Peter
