Can anyone offer any insights as to why SWISH-e running on an OS X box might
have either stalled or continues to run after two days? Any suggestions of
other tests I might run or settings I might tweak to correct this problem?
SWISH-e has been running economy mode - while top shows cpu activity (see
below), a gdb back trace yields the same report when checked at different
intervals (also below). The subdirectory listed in the gdb report
(/Library/Tomcat/blojsom_root/webapps/ROOT/blojsom_resources/meta/sdegraff )
contains 3 .mp3 files that are 10Mb, 3Mb and 2Mb.
SWISH-e has written 628 files with each filename beginning with swtmploc
followed by 6 or 7 other characters. It has not written the index file
(weblogs.swish). It has written
weblogs.swish.prop.temp (90936) and weblogs.swish.temp (421564) with the
last update yesterday afternoon.
Thanks!
------------------------
top:
Load Avg: 0.14, 0.09, 0.08 CPU usage: 0.9% user, 5.9% sys, 93.3% idle
SharedLibs: num = 80, resident = 2.84M code, 392K data, 588K LinkEdit
MemRegions: num = 18964, resident = 1.79G + 3.43M private, 7.47M shared
PhysMem: 194M wired, 1.19G active, 614M inactive, 1.98G used, 21.6M free
VM: 8.33G + 59.8M 68799545(1356) pageins, 5613058(4) pageouts
PID COMMAND %CPU TIME #TH #PRTS #MREGS
19074 swish-e 9.5% 25:12:14 1 13 7431
RPRVT RSHRD RSIZE VSIZE
1.66G- 1.14M 1.59G- 1.83G
gdb:
(gdb) bt
#0 0x90002b40 in strcmp ()
#1 0x000085c8 in getentry (sw=0x5bdcd374, word=0x1803400 "˻") at
index.c:1090
#2 0x00008de8 in addword (word=0x1803400 "˻", sw=0x1800400, filenum=1347,
structure=105, numMetaNames=1, metaID=0xc0, word_position=0xbffff784) at
index.c:2217
#3 0x0000ae68 in indexstring (sw=0x1800400, s=0xbfffdb5c "", filenum=1347,
structure=105, numMetaNames=1, metaID=0x1803800, position=0xbffff784) at
index.c:2525
#4 0x00016e6c in flush_buffer (parse_data=0xbffff738, clear=1902311) at
parser.c:1567
#5 0x00017c6c in start_metaTag (parse_data=0xbffff738, tag=0xbfffe4b9 "t",
endtag=0xbfffe4b9 "t", meta_append=0xbfffdcd8, prop_append=0xbfffdcdc,
is_html_tag=192) at parser.c:998
#6 0x000182dc in start_hndl (data=0xbffff738, el=0x1855336 "t", attr=0x0)
at parser.c:656
#7 0x92cc36b4 in htmlParseCharRef ()
#8 0x92cc5398 in htmlCreateMemoryParserCtxt ()
#9 0x92cc5c5c in htmlParseChunk ()
#10 0x00018540 in parse_chunks (parse_data=0xbffff738) at parser.c:338
#11 0x000186fc in parse_HTML (sw=0x1800400, fprop=0x31ea4, fi=0xbffff85c,
buffer=0x80808080 <Address 0x80808080 out of bounds>) at parser.c:256
#12 0x000099e0 in do_index_file (sw=0x0, fprop=0x41dc70) at index.c:961
#13 0x0000c238 in printfile (sw=0x1800400, filename=0x41dc70 "\001\230ü¼")
at fs.c:609
#14 0x0000c564 in indexadir (sw=0x1800400, dir=0x406f00
"/Library/Tomcat/blojsom_root/webapps/ROOT/blojsom_resources/meta/sdegraff")
at fs.c:649
#15 0x0000c5ec in indexadir (sw=0x1800400, dir=0x401e28
"/Library/Tomcat/blojsom_root/webapps/ROOT/blojsom_resources/meta") at
fs.c:687
#16 0x000035a8 in cmd_index (sw=0x1800400, params=0x33358) at swish.c:1388
#17 0x000045ac in main (argc=1541198708, argv=0xbffffca4) at swish.c:200
(gdb)
----- Original Message -----
Sent: Tuesday, October 17, 2006 11:38 AM
Subject: SWISH-e on OS X - not completing.
> Problem: SWISH-E does not seem to complete indexing large directory (2.1
> Gb, 252 subdirectories, mostly MP3, MP4, m4a files of 1 to 15Mb) on Mac OS
> X box.
>
> Works & completes normally on smaller directories with few subdirectories.
>
> Correction efforts: Combed mailing list archives & Googled. Found similar
> but different issues, some 5 years old, nothing to offer solution. Tried
> economy mode and regular mode.
>
> System specs:
>
> SWISH-E 2.4.3
> MAC OS X Server 10.4.7 (8jl35)
> Processor: 2 x 2.3 GHz
> Memory: 2048 MB
>
> Swish.conf:
>
> # Swish configuration file.
>
> # What directory to index
> IndexDir /Volumes/data/weblogs
> IndexDir
> /Library/Tomcat/blojsom_root/webapps/ROOT/blojsom_resources/meta
>
> # Output file
> IndexFile /Volumes/data/swish_index/weblogs.swish
>
> # What files to index
> IndexOnly .htm .html .txt .mp3
>
> # Tell Swish-e that .txt files are to use the text parser
> IndexContents TXT* .txt .meta
>
> # Otherwise, use the HTML parser
> DefaultContents HTML*
>
> # Replacing path info
> ReplaceRules replace "/Volumes/data/weblogs" "/weblog"
> ReplaceRules replace "/Library/Tomcat/blojsom_root/webapps/ROOT/" "/"
>
> Command: /usr/local/bin/swish-e -e -c
> /Volumes/data/swish_index/gpswish.conf
Received on Thu Oct 19 09:54:33 2006