| Anonymous | Login | Signup for a new account | 2013-05-22 17:42 CEST | ![]() |
| Main | My View | View Issues | Change Log | Roadmap |
| View Issue Details [ Jump to Notes ] | [ Issue History ] [ Print ] | ||||||
| ID | Project | Category | View Status | Date Submitted | Last Update | ||
| 0000617 | Sphinx | sphinxse | public | 2010-10-05 09:07 | 2012-09-29 12:04 | ||
| Reporter | kalron | ||||||
| Assigned To | Tomat | ||||||
| Priority | normal | Severity | crash | Reproducibility | always | ||
| Status | closed | Resolution | unable to reproduce | ||||
| Platform | OS | OS Version | |||||
| Product Version | 1.11-dev | ||||||
| Target Version | Fixed in Version | ||||||
| Summary | 0000617: Crashes on most queries | ||||||
| Description | My setup: - Index created using xmlpipe2 - Index size: 101GB - Tried both 1.10-beta as well as the latest release from SVN (r2500) - I'm using the Python API bindings to perform the searches - Currently used sphinx was compiled with: --without-mysql --with-debug For a few searches everything works as expected, but for the majority of queries searchd crashes. I haven't been able to identify a pattern to why certain queries are failing. Below is the output from one of the crashes: --- crashed SphinxAPI request dump --- AAABFwAAALAAAAABAAAAAAAAAAoAAAAGAAAAAAAAAAEAAAAJdGltZXN0YW1wAAAAD0BkYXRhIGNpc2NvIGlvcwAAAAAAAAAQc2hvZGFuX21vbmdvX2FsdAAAAAEAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAnEAAAAAtAZ3JvdXAgZGVzYwAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAABKg== --- crashed SphinxAPI request dump --- -------------- backtrace begins here --------------- Sphinx 1.11-dev (r2500) Program compiled with gcc 4.4.3 Host OS is Linux 2.6.32-24-server 0000039-Ubuntu SMP Wed Jul 28 06:21:40 UTC 2010 x86_64 GNU/Linux Stack bottom = 0x7fffe1f5a21f, thread stack size = 0x14000 begin of manual backtrace: Something wrong with thread stack, backtrace may be incorrect (fp=7fffe1f53e10) Stack looks OK, attempting backtrace. 406051 7f2648786a75 Something wrong in frame pointers, backtrace failed (fp=7f2648899ada) begin of system backtrace: begin of system symbols: ./searchd[0x50b1fc] ./searchd[0x406051] /lib/libpthread.so.0(+0xf8f0)[0x7f26492938f0] /lib/libc.so.6(gsignal+0x35)[0x7f2648786a75] /lib/libc.so.6(abort+0x180)[0x7f264878a5c0] /lib/libc.so.6(__assert_fail+0xf1)[0x7f264877f941] ./searchd[0x4d1512] ./searchd[0x496fe0] ./searchd[0x4d25a4] ./searchd[0x46f5d1] ./searchd[0x54f2ea] ./searchd[0x54f765] ./searchd[0x54fb98] ./searchd[0x5500ad] ./searchd[0x5594f6] ./searchd[0x564f4a] ./searchd[0x55c587] ./searchd[0x478233] ./searchd[0x47715d] ./searchd[0x477325] ./searchd[0x412320] ./searchd[0x413aec] ./searchd[0x410611] ./searchd[0x4154e1] ./searchd[0x42116c] ./searchd[0x426255] ./searchd[0x42e57d] ./searchd[0x431fb2] ./searchd[0x4320e5] /lib/libc.so.6(__libc_start_main+0xfd)[0x7f2648771c4d] ./searchd[0x403f99] -------------- backtrace ends here --------------- | ||||||
| Additional Information | I rebuilt the index twice, just in case something went wrong the first time. I wonder whether the crashes have something to do with the index size. An older index I'm currently using isn't experiencing any of these problems. | ||||||
| Tags | No tags attached. | ||||||
| Attached Files | |||||||
Notes |
|
|
(0000914) kalron (reporter) 2010-10-08 19:21 |
Got this error now when trying to use the indexer: indexer: sphinx.cpp:16401: virtual void CSphDictCRCTraits::DictEntry(SphWordID_t, BYTE*, int, int, SphOffset_t, SphOffset_t): Assertion `m_wrDict.GetPos()<=(2147483647 * 2U + 1U)' failed. *** Oops, indexer crashed! Please send the following report to developers. -------------- report begins here --------------- Current document: docid=44509981, hits=28359551 Current batch: minid=0, maxid=0 Hit pool start: docid=42723603, hit=78293768 -------------- backtrace begins here --------------- Sphinx 1.11-dev (r2500) Program compiled with gcc 4.4.3 Host OS is Linux heimdal 2.6.32-24-server 0000039-Ubuntu SMP Wed Jul 28 06:21:40 UTC 2010 x86_64 GNU/Linux Stack bottom = 0x0, thread stack size = 0x14000 begin of system backtrace: begin of system symbols: ./indexer[0x4cf190] ./indexer[0x409d86] /lib/libpthread.so.0(+0xf8f0)[0x7fcfd86a48f0] /lib/libc.so.6(gsignal+0x35)[0x7fcfd7b97a75] /lib/libc.so.6(abort+0x180)[0x7fcfd7b9b5c0] /lib/libc.so.6(__assert_fail+0xf1)[0x7fcfd7b90941] ./indexer[0x448d55] ./indexer[0x4264e0] ./indexer[0x430d0c] ./indexer[0x408942] ./indexer[0x40b4d0] /lib/libc.so.6(__libc_start_main+0xfd)[0x7fcfd7b82c4d] ./indexer[0x403499] ---------------------------------- What does the assertion error mean? |
|
(0000919) Tomat (manager) 2010-10-18 20:53 |
Could you post binaries either search daemon and indexer and sym files from these crashes? I was wandering about index integrity. Could you check it with this command ./indextool -c YOUR_CONFIG --check YOUR_INDEX_NAME and post result here. |
Issue History |
|||
| Date Modified | Username | Field | Change |
| 2010-10-05 09:07 | kalron | New Issue | |
| 2010-10-05 09:07 | kalron | File Added: searchd.sym | |
| 2010-10-08 19:21 | kalron | Note Added: 0000914 | |
| 2010-10-18 20:53 | Tomat | Note Added: 0000919 | |
| 2010-10-18 20:53 | Tomat | Assigned To | => Tomat |
| 2010-10-18 20:53 | Tomat | Status | new => assigned |
| 2012-09-29 12:04 | kevg | Status | assigned => closed |
| 2012-09-29 12:04 | kevg | Resolution | open => unable to reproduce |
| Copyright © 2000 - 2010 MantisBT Group |




