Fri Mar 13 15:51:04 PDT 2009
- Previous message: [Slony1-bugs] [Bug 77] New: Missing confirm records for events in sl_confirm table
- Next message: [Slony1-bugs] [Bug 78] New: Vacuum of sl_log_2 with 42M rows is blocking replication
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
http://www.slony.info/bugzilla/show_bug.cgi?id=71 César Acuña <cjacuna at gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |cjacuna at gmail.com --- Comment #3 from César Acuña <cjacuna at gmail.com> 2009-03-13 15:51:04 --- Chris... sorry for the late response. I managed to narrow the problem to the function void dump_configuration(void) in src/slon/confoptions.c void dump_configuration(void) { int i; for (i = 0; ConfigureNamesInt[i].gen.name; i++) { slon_log(SLON_CONFIG, "main: Integer option %s = %d\n", ConfigureNamesInt[i].gen.name, *(ConfigureNamesInt[i].variable)); } for (i = 0; ConfigureNamesBool[i].gen.name; i++) { slon_log(SLON_CONFIG, "main: Boolean option %s = %d\n", ConfigureNamesBool[i].gen.name, *(ConfigureNamesBool[i].variable)); } for (i = 0; ConfigureNamesReal[i].gen.name; i++) { slon_log(SLON_CONFIG, "main: Real option %s = %f\n", ConfigureNamesReal[i].gen.name, *(ConfigureNamesReal[i].variable)); } for (i = 0; ConfigureNamesString[i].gen.name; i++) { --> Here //slon_log(SLON_CONFIG, "main: String option %s = %s\n", ConfigureNamesString[i].gen.name, *(ConfigureNamesString[i].variable)); } } I commented out the line slon_log(SLON_CONFIG, "main: String option %s = %s\n", ConfigureNamesString[i].gen.name, *(ConfigureNamesString[i].variable)); and everything works fine. Doing a little debugging, I found out that the problem arises when it tries to dump to the log the content of the following configuration variables: pid_file archive_dir sql_on_connection lag_interval command_on_logarchive I appreciate the time you devoted to this problem. Greetings, César Acuña (In reply to comment #2) > Do you get any output before the crash, or is it buffered, and therefore lost? > > That crash takes place while it's logging the set of configuration parameters > to the slon log (file, or standard output). > > If you have a way to capture the output in an unbuffered fashion, that would be > pretty useful, as that would let us get closer to finding where it fails. > > *Apparently*, as it's trying to dump the config parameters, one of them is > failing to be accessed properly, hence the seg fault. Perhaps one of them > isn't properly enough initialized, and it would seem a mighty good thing if we > can figure out at what point the failure takes place. > > I'm not so familiar with Solaris; can you submit output to an unbuffered > destination so that you capture as much log as possible? If you can do that, > that should make it pretty easy to track down where the problem would take > place. > -- Configure bugmail: http://www.slony.info/bugzilla/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. You are the assignee for the bug.
- Previous message: [Slony1-bugs] [Bug 77] New: Missing confirm records for events in sl_confirm table
- Next message: [Slony1-bugs] [Bug 78] New: Vacuum of sl_log_2 with 42M rows is blocking replication
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Slony1-bugs mailing list