Mantis - ISSUES

0000591: Random results with SPH_SORT_EXPR mode on SetOverriden attributes.

The following used to work fine on Sphinx 0.9.9 (r2117).

import sphinxapi
import config_sphinx

def make_client():
cl = sphinxapi.SphinxClient()
cl.SetServer('localhost', config_sphinx.port)
return cl

def test_set_override(cl):
year = {111111:1, 111112:10, 111113:100}
cl.SetMatchMode(sphinxapi.SPH_MATCH_EXTENDED2)
cl.SetOverride('year_attr', sphinxapi.SPH_ATTR_INTEGER, year)
#cl.SetSortMode(sphinxapi.SPH_SORT_EXTENDED, 'year_attr desc')
cl.SetSortMode(sphinxapi.SPH_SORT_EXPR, 'year_attr')
return cl.Query('111111 | 111112 | 111113')

cl = make_client()
results = test_set_override(cl)

for r in results['matches']:
print r['id'], r['attrs']['year_attr']

with results:

>> 111112 10
>> 111111 1
>> 111113 100

However if the mode is set to extended we get the expected results:

>> 111113 100
>> 111112 10
>> 111111 1

0000590: Can't have a field named "status"

In my Sphinx configuration I have the following:

sql_attr_uint = status

This query:

SELECT * FROM nolimits WHERE status = 1 ORDER BY rating DESC OPTION ranker=bm25;

produces this error:

ERROR 1064 (42000): syntax error, unexpected STATUS, expecting IDENT or ID or MATCH near 'status = 1 ORDER BY rating DESC OPTION ranker=bm25'

0000589: LIBEXPAT.dll is not included in sphinx-1.10-beta-win32-id64-full.zip

LIBEXPAT.dll is not included in sphinx-1.10-beta-win32-id64-full.zip and when we want to run indexer.exe windows will alert the following message:
The program can't start because LIBEXPAT.dll is missing from your computer.

I attached this file to be placed in bin directory.
Thank you.

0000588: Sphinx segmentation fault when searching

i have a problem with index - sphinx fail when try search
it gives me an error message "Segmentation fault"

./search -c ../etc/sphinx.conf -i rts new
?????? ???????????????

Then i try to check index with indextool

Sphinx 1.11-id64-dev (r2472)
Copyright (c) 2001-2010, Andrew Aksyonoff
Copyright (c) 2008-2010, Sphinx Technologies Inc ( http://sphinxsearch.com [^] )

using config file '../etc/sphinx.conf'...
checking index 'rts'...
checking dictionary...
checking data...
FAILED, row not found (wordid=1166983373492(??A?), docid=388885)
FAILED, row not found (wordid=26815325588045(??A?), docid=622244)
FAILED, row not found (wordid=28743421987717(??A?), docid=427672)
FAILED, row not found (wordid=30061363013751(??A?), docid=1212197)
FAILED, row not found (wordid=30061363013751(??A?), docid=1253906)
FAILED, row not found (wordid=30810437118623(??A?), docid=327840)
FAILED, row not found (wordid=33880271985906(??A?), docid=53040)
FAILED, row not found (wordid=35489705271929(??A?), docid=76331)
FAILED, row not found (wordid=36912080341550(??A?), docid=390839)
FAILED, row not found (wordid=36912080341550(??A?), docid=404816)
FAILED, row not found (wordid=36912080341550(??A?), docid=415335)
FAILED, row not found (wordid=36912080341550(??A?), docid=415841)
FAILED, row not found (wordid=36912080341550(??A?), docid=415890)
FAILED, row not found (wordid=36912080341550(??A?), docid=416860)
FAILED, row not found (wordid=36912080341550(??A?), docid=417919)
FAILED, row not found (wordid=36912080341550(??A?), docid=418504)
FAILED, row not found (wordid=36912080341550(??A?), docid=419705)
FAILED, row not found (wordid=36912080341550(??A?), docid=507689)
FAILED, row not found (wordid=36912080341550(??A?), docid=508107)
FAILED, row not found (wordid=36912080341550(??A?), docid=511565)
FAILED, row not found (wordid=36912080341550(??A?), docid=511575)
FAILED, row not found (wordid=36912080341550(??A?), docid=511951)
FAILED, row not found (wordid=36912080341550(??A?), docid=512012)
FAILED, row not found (wordid=36912080341550(??A?), docid=512961)
FAILED, row not found (wordid=36912080341550(??A?), docid=512990)
FAILED, row not found (wordid=36912080341550(??A?), docid=514710)
FAILED, row not found (wordid=36912080341550(??A?), docid=514760)
FAILED, row not found (wordid=36912080341550(??A?), docid=1132076)
FAILED, row not found (wordid=36912080341550(??A?), docid=1132078)
FAILED, row not found (wordid=39186036870579(??A?), docid=571412)
FAILED, row not found (wordid=39186036870579(??A?), docid=574938)
FAILED, row not found (wordid=39186036870579(??A?), docid=577836)
FAILED, row not found (wordid=39186036870579(??A?), docid=723682)
FAILED, row not found (wordid=41187820428553(??A?), docid=1126177)
FAILED, row not found (wordid=45897058083512(??A?), docid=622569)
FAILED, row not found (wordid=45897058083512(??A?), docid=624095)
FAILED, row not found (wordid=45897058083512(??A?), docid=624324)
FAILED, row not found (wordid=46087447849255(??A?), docid=151895)
FAILED, row not found (wordid=48683844810461(??A?), docid=399115)
FAILED, row not found (wordid=48683844810461(??A?), docid=442822)
FAILED, row not found (wordid=48683844810461(??A?), docid=928677)
FAILED, row not found (wordid=48683844810461(??A?), docid=1036223)
FAILED, row not found (wordid=48829807817615(??A?), docid=325119)
FAILED, row not found (wordid=50105661882079(??A?), docid=462733)
FAILED, row not found (wordid=60583269066933(??A?), docid=660669)
FAILED, row not found (wordid=60583269066933(??A?), docid=704258)
FAILED, row not found (wordid=60583269066933(??A?), docid=705032)
FAILED, row not found (wordid=61167342978675(??A?), docid=860031)
FAILED, row not found (wordid=70849543353948(??A?), docid=383937)
FAILED, row not found (wordid=74163257211519(??A?), docid=1117102)
FAILED, row not found (wordid=75072123887525(??A?), docid=76331)
FAILED, row not found (wordid=76850467087326(??A?), docid=602237)
FAILED, row not found (wordid=76850467087326(??A?), docid=884312)
FAILED, row not found (wordid=83280564280876(??A?), docid=714441)
FAILED, row not found (wordid=86157596528283(??A?), docid=78434)
FAILED, row not found (wordid=86453634932553(??A?), docid=678654)
FAILED, row not found (wordid=91821485348497(??A?), docid=1129861)
FAILED, row not found (wordid=95433507507615(??A?), docid=980526)
FAILED, row not found (wordid=95977988057691(??A?), docid=522263)
FAILED, row not found (wordid=95977988057691(??A?), docid=522267)
FAILED, row not found (wordid=95977988057691(??A?), docid=522269)
FAILED, row not found (wordid=95977988057691(??A?), docid=597373)
FAILED, row not found (wordid=95977988057691(??A?), docid=597377)
FAILED, row not found (wordid=95977988057691(??A?), docid=681590)
FAILED, row not found (wordid=95977988057691(??A?), docid=681598)
FAILED, row not found (wordid=95977988057691(??A?), docid=681599)
FAILED, row not found (wordid=95977988057691(??A?), docid=681600)
FAILED, row not found (wordid=95977988057691(??A?), docid=702003)
FAILED, row not found (wordid=95977988057691(??A?), docid=702010)
FAILED, row not found (wordid=95977988057691(??A?), docid=702026)
FAILED, row not found (wordid=102809208949814(??A?), docid=535982)
FAILED, row not found (wordid=105973893850577(??A?), docid=881393)
FAILED, row not found (wordid=108844449880737(??A?), docid=900959)
FAILED, row not found (wordid=108844449880737(??A?), docid=903764)
FAILED, row not found (wordid=121787539880377(??A?), docid=161407)
FAILED, row not found (wordid=123550946352916(??A?), docid=796052)
FAILED, row not found (wordid=124862979748062(??A?), docid=70969)
FAILED, row not found (wordid=125848661401527(??A?), docid=397684)
FAILED, row not found (wordid=128915087522209(??A?), docid=393821)
FAILED, row not found (wordid=128915087522209(??A?), docid=402289)
FAILED, row not found (wordid=128915087522209(??A?), docid=405592)
FAILED, row not found (wordid=128915087522209(??A?), docid=415211)
FAILED, row not found (wordid=128915087522209(??A?), docid=442006)
FAILED, row not found (wordid=128915087522209(??A?), docid=443531)
FAILED, row not found (wordid=128915087522209(??A?), docid=450484)
FAILED, row not found (wordid=128915087522209(??A?), docid=472745)
FAILED, row not found (wordid=131768785336887(??A?), docid=4280)
FAILED, row not found (wordid=139894673232382(??A?), docid=1185371)
FAILED, row not found (wordid=139894673232382(??A?), docid=1204967)
FAILED, row not found (wordid=139894673232382(??A?), docid=1211483)
FAILED, row not found (wordid=139894673232382(??A?), docid=1211484)
FAILED, row not found (wordid=139894673232382(??A?), docid=1223979)
FAILED, row not found (wordid=139894673232382(??A?), docid=1223980)
FAILED, row not found (wordid=145541178917136(??A?), docid=755872)
FAILED, row not found (wordid=145541178917136(??A?), docid=871560)
FAILED, row not found (wordid=146473578994454(??A?), docid=622538)
FAILED, row not found (wordid=146473578994454(??A?), docid=622833)
FAILED, row not found (wordid=151062713822239(??A?), docid=333485)
FAILED, row not found (wordid=153727707578801(??A?), docid=418257)
FAILED, row not found (wordid=157659783591505(??A?), docid=866840)
(threshold reached; suppressing further output)
checking rows...
checking attribute blocks index...
checking kill-list...
check FAILED, 100 of 243206647 failures reported, 72.2 sec elapsed

It write that he could not find something.

And it is crash log:

-------FATAL: CRASH LOG DUMP-------
[Tue Aug 31 14:58:22.126 2010] [21385]

--- crashed SphinxAPI request dump ---
AAABFwAAAOYAAAABAAAAAAAAAAoAAAAGAAAAAAAAAAAAAAAAAAAAGEAodGl0bGUsYW5jaG9yKSAid2VsbCIvMQAAAA
AAAAANcnRzLCBydHNEZWx0YQAAAAEAAAAAAAAAAAAAAAAAAAAAAAAAAQAAAAZraWxsZWQAAAAAAAAAAQAAAAAAAAAA
AAAAAAAAAAAAAAAAAAAACgAAAAtAZ3JvdXAgZGVzYwAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAACAAAABX
RpdGxlAAAACgAAAAZhbmNob3IAAAAUAAAAAAAAAAAAAAABKg==
--- crashed SphinxAPI request dump ---
-------------- backtrace begins here ---------------
Sphinx 1.11-id64-dev (r2472)
Program compiled with gcc 4.4.3
Host OS is Linux us1 2.6.32-24-server 0000041-Ubuntu SMP Thu Aug 19 02:47:08 UTC 2010 x86_64
GNU/Linux
Stack bottom = 0x7fe99b93ee7f, thread stack size = 0x14000
begin of manual backtrace:
Something wrong with thread stack, backtrace may be incorrect (fp=7fe99b93e090)
Stack looks OK, attempting backtrace.
4119d6
54cd2e
46f248
54f5d2
563099
4715f4
4973b5
49adef
49a53f
422184
4223d5
50f27d
7fe99b2579ca
Something wrong in frame pointers, backtrace failed (fp=0)
begin of system backtrace:
begin of system symbols:
/usr/local/sphinx/bin/searchd[0x508ac4]
/usr/local/sphinx/bin/searchd(_ZN16SphCrashLogger_c11HandleCrashEi+0x146)[0x4119d6]
/lib/libpthread.so.0(+0xf8f0)[0x7fe99b2608f0]
/usr/local/sphinx/bin/searchd[0x54cd2e]
/usr/local/sphinx/bin/searchd[0x46f248]
/usr/local/sphinx/bin/searchd[0x54f5d2]
/usr/local/sphinx/bin/searchd[0x563099]
/usr/local/sphinx/bin/searchd[0x4715f4]
/usr/local/sphinx/bin/searchd[0x4973b5]
/usr/local/sphinx/bin/searchd[0x49adef]
/usr/local/sphinx/bin/searchd[0x49a53f]
/usr/local/sphinx/bin/searchd[0x422184]
/usr/local/sphinx/bin/searchd(_Z21LocalSearchThreadFuncPv+0x45)[0x4223d5]
/usr/local/sphinx/bin/searchd(_Z20sphThreadProcWrapperPv+0x1d)[0x50f27d]
/lib/libpthread.so.0(+0x69ca)[0x7fe99b2579ca]
/lib/libc.so.6(clone+0x6d)[0x7fe99a3e76fd]
-------------- backtrace ends here ---------------

I don't understand what is happen: early it worked with the index normally, and now
writes error.
And in the configuration files do not fundamentally changed.
ReIndex not help.

What happened and how to fix this?
Can a source table size affect on it?

0000587: filtering with python api lib disallows integers values bigger than 0x7fffffff

In the code of SetFilter(api/sphinxapi.py):
=====================
for value in values:
assert(isinstance(value, int))
=====================
int type in python only corresponds to int32_t type in c and if an integer larger than 0x7fffffff is used it will always be automatically converted to long so assertion will fail.

suggesting to change this into
assert(isinstance(value, (int, long)))
or more strict by checking the integer range.

0000586: SphinxQL erratically returns NULL result, but no associated mysql_error

I'm connecting to Sphinx via SphinxQL from PHP. Every so often, Sphinx returns NULL as a result of two certain queries, but mysql_error() returns empty string. Re-running the same query again works ok. The examples of the failing queries (see the "Additional Information" for the index definition):

> SELECT * FROM posts WHERE userid = 325249 GROUP BY forumid ORDER BY @count DESC LIMIT 1

> SELECT * FROM posts WHERE userid = 45231 GROUP BY threadid ORDER BY @count DESC LIMIT 1

The server where Sphinx is running isn't loaded at all, the query rate is about one in 1-5 seconds. There is nothing in searchd.log, and I was able to track down the associated entries in query.log - they look ok, there is nothing unusual:

> [Fri Aug 20 12:55:28.923 2010] 0.106 sec [ext2/1/ext 0 (0,1) @forumid] [posts] [ios=0 kb=0.0 ioms=0.0 cpums=191.6]

> [Fri Aug 20 12:41:01.557 2010] 0.421 sec [ext2/1/ext 5 (0,1) @threadid] [posts] [ios=0 kb=0.0 ioms=0.0 cpums=504.8]

The failing queries are always of the same two types outlined above, but the userid parameter is different each time. Again, re-running a failed query again, either manually or by visiting the page that triggers the query works ok. The failure rate is quite low - for example, there were 12 failures in a day out of 3672 queries of that type.

There are many more different queries against the same index such as other scan queries and the fulltext queries, but they always return the results (no NULL as a result). Though most of them are being executed through the PHP API, but there are some SphinxQL queries as well.

Sometimes there is a MySQL error returned by SphinxQL: "Server shutdown in progress". It only happens during the index rotation. Again, only this particular type of query returns this MySQL error - I didn't see "server shutdown in progress" ever before.

I'm currently using 1.10beta, but the same errors were happening in 0.9.9.

0000585: str2ordinal attributes are always 0 if docinfo is set to inline

str2ordinal attributes are always 0 if docinfo is set to inline, when set back to extern, it's working normally again.
Maybe it's a documentation bug or probably a program bug since the storagement sounds the same for uint and ordinals.
Please take a look at it. Thanks.

Peace.

0000584: Bug: broken SetOverride in distributed index

Hi all,

It seems the newest version of Sphinx (1.10-beta) has broken SetOverride when using
distributed indexes. Is anyone experiencing these difficulties?

cl = sphinxapi.SphinxClient()
cl.SetServer('localhost', port)
cl.SetOverride('score_attr', sphinxapi.SPH_ATTR_FLOAT, scores)
cl.Query('hello') # same problem even with cl.Query('hello', index='dist_index')

> error: unpack str size does not match format

and sphinx.conf file is everything you would expect.

I tried everything so I really think there might be a bug here. If anyone has it working
please let me know.

Thank you,

Alex

0000583: Indexer crashes at use wordforms

Indexer crashes on certain words from wordforms file.

File of wordforms contents russian words in utf8 encoding.

0000542: IN Clause fails to find all documents

I use IN clause for querying the sphinx, using:
SELECT * FROM entries WHERE entry_status IN(101,102,0,1,2) AND media_type IN(1,2,5,6,201) AND partner_id = 291 LIMIT 20;

In the index I have 11 documents in entry_status 2 and one in entry_status 1 and yet I get only the first 11 results.

However, when I query only for statuses 1 and 2 I get all 12 results, as followed:
SELECT * FROM entries WHERE entry_status IN(1,2) AND media_type IN(1,2,5,6,201) AND partner_id = 291 LIMIT 20;

0000582: Fatal Crash in searchd

From the backtrace, it looks like there is a fatal crash when filtering, although I have not looked into it too deep. It is reproducable using the same request that was dumped.

0000485: Wrong (additional) results when rotating

Its actually sphinx 0.9.9-release (r2117)

I have two indices (main and delta). Now, as documents become added, I run an "indexer --rotate delta". Search is done over both indices.

Now I sometimes get results which actually are none. They are not related at all. But most wrong results have a id (auto inc in mysql) close together.

I noticed it, because I have a rss feed to check for new results and my reader is archiving all results.
Knowing this I made script, always doing the same query and waiting for a change in the results. And so I got this example:

ids ok: 62192152, 61614749, 61169581, 60865306, 57267807

wrong additional ids in one search:
63297709, 63297710, 63297711, 63297712, 63297713,
63297721,
63297733,
63297779, 63297780, 63297781, 63297782, 63297783, 63297784,
63297714, 63297715, 63297716, 63297717, 63297718

Database and index is far too big to upload. But it should be reproduceable with smaller datasets.

0000478: SetGroupBy and filter by @count

Example of output data without filter by @count
row1:([id]=>1,[file_id]=>45,[keyword_id]=>51,[@groupby]=>1,[@count]=>4)
row2:([id]=>2,[file_id]=>46,[keyword_id]=>50,[@groupby]=>2,[@count]=>5)

Trying to filter data by @count:
Way 1:
$cl->SetSelect( "*,@count as group_count");
-> "parse error: unknown identifier '@count' (not an attribute, not a function) "
Way 2: $cl->SetFilter('@count',array(5)); -> filtering all, result is empty
Way 3: $cl->SetMatchMode(SPH_MATCH_EXTENDED); $cl->Query('@count 5',$index);
-> query error: no field 'count' found in schema

So @count can't be renamed and @count can't be used in SetFilter or Query.

0000581: geodistance optimization with other attribute filters

Sphinx would calculate geo distance for all attributes even if there are other filters that would potentially limit the amount of data to process significantly. Here's one example with geofilter and without it:

/usr/local/src/sphinx-0.9.9-rc2/api# php -q test-geo.php -h 127.0.0.1 -i location -fr created_at 1283008618 1283181418 -l 1
Query '' retrieved 7 of 7 matches in 0.245 sec.
Query stats:

Matches:
1. doc_id=22, weight=1, image_id=3, lat_rad=0.746088087559, long_rad=-1.37429499626, created_at=2010-08-28 16:17:27, @geodist=19308.6054688

/usr/local/src/sphinx-0.9.9-rc2/api# php -q test.php -h 127.0.0.1 -i location -fr created_at 1283008618 1283181418 -l 1
Query '' retrieved 1000 of 39455 matches in 0.032 sec.
Query stats:

Matches:

1. doc_id=12, weight=1, image_id=6, lat_rad=0.795294225216, long_rad=0.103550389409, created_at=2010-08-28 11:17:00

0000146: Mysql replication and Sphinx

sphix is really great, but I experience a very strange situation with my sphinx server.
So I have master-slave scheme and in the slave I compiled the sphinx engine in mysql 5.0.37, to use it with joins with one of my tables, which is replicating from the master (in fact all tables are raplicating, but I use only two - the indexer use one and then the query uses sphinx table joining one more table). The replication is 300-400 queris/sec and when the replication is on, sphinx is so fast taht there is a hardly one or two queries shown in "show processlist" from time to time, but after a while (1-5 minutes) suddenly some queries (a small amount at first 5-10 queries) just hang in "Sending data", and after that almost all queries get "Locked" state, the others get hung in "Sending data" and the max_connection is reached for seconds. In the meawhile the load of the server os almost 0 (no load - 0.03-0.20), no IO, nothing. The searchd daemon does nothing (straced it, and the daemon just listens for new connections).
I cannot figure it out what makes mysql suddenly to "Sending data" and not getting out of this state despite of there is no server load or reindexing!

0000558: Feature request: syslog logging

In order to support logging to another system, the possibility of logging to syslog would be nice to have. Alternatively, just logging to an external process (through a pipe) would be sufficient, such that another tool can log to syslog (this would give more flexibility to configuring facility, program, log level, etc as I already have my own script for this purpose).

0000567: Able to return term position in the search result

It would be useful to return the term position(s) in the search result. So we can do some post-processing on the search result.

0000580: SphinxQL: query output does not retain column order

When I query the index, response does not match the order of columns that were originally in the query. Here's an example:

mysql> SELECT user_id, nomeabrev, estado, categ_id, @count from index where match('@(key) something') group by user_id;
+----------+--------+---------+---------------------+--------------------------+----------+--------+
| id | weight | user_id | estado | nomeabrev | categ_id | @count |
+----------+--------+---------+---------------------+--------------------------+----------+--------+
| 1 | 1697 | 1 | Goiás | Some user | 2 | 4 |
| 8 | 1691 | 9 | Paraíba | Other user | 2 | 4 |
| 12 | 1691 | 5 | Rio Grande do Sul | User 0000005 | 1 | 1 |
| 29 | 1691 | 2 | Rio de Janeiro | Last User here | 3 | 7 |

You can see estado and nomeabrev get reversed. Seems like a bug to me.

0000579: ODBC indexing crash, 32/64 bit issue; source code error spotted

Take a peak at:
http://code.google.com/p/sphinxsearch/source/browse/trunk/src/sphinx.cpp#22284 [^]

It is not difficult to see that a line there is duplicated.

What I experience;
In the ODBC driver the *nullp pointer is basically made zero on the msb side of the integer.

This introduces:
Hardware watchpoint 1: ((ODBCDescRec *) 0x7a61f0)->sql_desc_indicator_ptr

Old value = (SQLLEN *) 0x79f978
New value = (SQLLEN *) 0xf978
CSphSource_ODBC::SqlFetchRow (this=0x76b920) at sphinx.cpp:22277
22277 ARRAY_FOREACH ( i, m_dColumns )

I guess the 'idea' here is to make the stored value zero. This is not what is actually happening. Now I wonder several things.

Please confirm:
1) tCol.m_dContents stores the pointer to the data in the ODBC records
2) tCol.m_dContents should shore a value of size SQLLEN and not char.

0000578: BuildExcerpts query_mode=true crashes searchd

If you set query_mode to true and run BuildExcerpts, searchd crashes. :(

0000524: Random Crash

I am running Sphinx 1.10.1-dev (r2320) and getting some random crashes.

Attached are both crash files and the log file as well. You will see
in the log file I am sending a SIGHUP every 10 minutes. This is to
"force" Sphinx to write the realtime index data to file, so when it
crashes, it does not lose the realtime index completely. I figured
this would be the best method until binary logging/recovery is added.

I am not sure if the SIGHUP has anything to do with the crash.. I expect it is something completely different since it never had crashed until a few minutes after issuing SIGHUP to searchd. It must have something to do with a query, even though the queries I was testing were all alike.

0000518: MySQL times out while indexing and aborts indexer.

When indexing large amount of data(Note: 0000594 935 116 rows) indexer loses it's connection to MySQL. Instead of reconnecting, indexer aborts and I have to start the indexing from scratch. Since this ALWAYS occurs when I'm about half way through(2-3 hours of indexing) I'm not able to complete this index.

I've attempted to patch it myself with adding MYSQL_OPT_RECONNECT to the MySQL connect-function, but it does not solve the problem.

Error message is:
ERROR: index 'messages': sql_fetch_row: Lost connection to MySQL server during query.

0000577: Disk space never reclaimed when using RT (real-time) indexes

The RT indexes only grow in size even when REPLACE INTO is used to insert exactly the same data, or a DELETE FROM is used to delete all rows one by one before reinserting them.

Here's the disk space usage after the insertion of about 700K rows:
> du -sk /usr/local/var/sphinx_rt/
387740 /usr/local/var/sphinx_rt/

Disk usage after replacing the same data with REPLACE INTO:
> du -sk /usr/local/var/sphinx_rt/
811188 /usr/local/var/sphinx_rt/

Disk usage after (supposedly) deleting all data with DELETE FROM, one by one row, and exiting searchd to flush the RAM data:
> du -sk /usr/local/var/sphinx_rt/
811908 /usr/local/var/sphinx_rt/

Disk usage after re-inserting all data again and shutting down searchd:
> du -sk /usr/local/var/sphinx_rt/
1234108 /usr/local/var/sphinx_rt/

0000557: SphinxQL: QUOTED_STRING not parsed correctly

The following queries are parsed successfully:

SELECT * FROM index WHERE MATCH('abc xyz');
SELECT * FROM index WHERE MATCH('"abc xyz"');

But the following queries yield a syntax error (unexpected $end near '')

SELECT * FROM index WHERE MATCH('"abc xyz');
SELECT * FROM index WHERE MATCH('"abc xyz" "');

In general, un uneven number of quotes seems to cause a syntax error.

0000576: libsphinxclient improperly handling max=0 in SetLimits()

This error duplicate of 0000208 but I have same error with Sphinx 0.9.9 (r2117; Dec 02, 2009) version.

When I use SetLimits(0,10) sphinx give this error:
per-query max_matches=0 out of bounds (per-server max_matches=1000)