On Tue, Mar 31, 2009 at 6:21 PM, Peter Hsu <[email protected]> wrote: > I got a couple of traces like that... they all looked the same. Couldn't > get any other tracebacks. Any other ideas? About to move on to a newer OS, > though I am severely disappointed by this one. >
Other than adding a debug statement in the second definition of couch_os_process:terminate/2 when Port is not nil and writing out that its being closed and the reason I got nothing. Reading through the couch_query_servers.erl I don't see anywhere that couch would be proactively closing the process which jives with my memory of the situation. Perhaps the EPIPE is a red herring and something else is getting whacked? If you're looking for suggestions, Ubuntu 8.04 or 8.10 is definitely at the top of my list. I still have a sneaking suspicion that its an improperly built spidermonkey but that's mostly because you said you built from source and building spidermonkey from source is like walking through a minefield. Sometimes you get across and other times, well, not so much. HTH, Paul Davis > On Mar 31, 2009, at 12:02 PM, Paul Davis wrote: > >> Odd. Erroring out on a broken pipe like that sure would seem like >> erlang is closing the process. >> >> Is there by chance more traceback that looks slightly different? I'd >> expect some sort of other bad match error earlier on if erlang is >> deciding to kill the process. >> >> Paul >> >> On Tue, Mar 31, 2009 at 2:46 PM, Peter Hsu <[email protected]> wrote: >>> >>> Program received signal SIGPIPE, Broken pipe. >>> [Switching to Thread 182897363104 (LWP 30503)] >>> 0x0000003c051b9f12 in __write_nocancel () from /lib64/tls/libc.so.6 >>> (gdb) where >>> #0 0x0000003c051b9f12 in __write_nocancel () from /lib64/tls/libc.so.6 >>> #1 0x0000003c05165678 in _IO_new_file_write () from /lib64/tls/libc.so.6 >>> #2 0x0000003c051646b4 in _IO_new_do_write () from /lib64/tls/libc.so.6 >>> #3 0x0000003c05164ffe in _IO_new_file_sync () from /lib64/tls/libc.so.6 >>> #4 0x0000003c0515a930 in fflush () from /lib64/tls/libc.so.6 >>> #5 0x0000000000402042 in Print (context=0x50b0b0, obj=Variable "obj" is >>> not >>> available. >>> ) at couch_js.c:275 >>> #6 0x0000002a95709bc9 in js_Invoke () from /usr/local/lib/libjs.so >>> #7 0x0000002a95713271 in js_Interpret () from /usr/local/lib/libjs.so >>> #8 0x0000002a9570a2c4 in js_Execute () from /usr/local/lib/libjs.so >>> #9 0x0000002a956e3498 in JS_ExecuteScript () from >>> /usr/local/lib/libjs.so >>> #10 0x0000000000403979 in main (argc=Variable "argc" is not available. >>> ) at couch_js.c:391 >>> (gdb) >>> >>> Spider monkey was installed from 1.7.0 source. >>> >>> On Mar 31, 2009, at 11:40 AM, Paul Davis wrote: >>> >>>> On Tue, Mar 31, 2009 at 2:26 PM, Peter Hsu <[email protected]> wrote: >>>>> >>>>> Don't think so: >>>>> >>>>> [r...@rhel4-dell apache-couchdb-0.9.0]# ldd /usr/local/bin/couchjs >>>>> not a dynamic executable >>>>> >>>>> Is there any way to get more visibility into the problem? Debug >>>>> logging? >>>>> >>>> >>>> Hmm, debug logging probably won't get you much beyond the erlang >>>> traceback. IIRC, there's no debug option for couchjs. >>>> >>>> AFAICT without adding debug statements to your setup is that the error >>>> is happening when couch_query_server sends ["reset"] to the couchjs >>>> process. From my initial poking through the various bits it looks like >>>> couchjs is crashing as opposed to exiting cleanly. A quick way to >>>> diagnose this is to start up couchdb, get a couchjs process started, >>>> attach to it using gdb, and then trigger the error to see if you can >>>> get a traceback. That sounds not quick now that I typed it out. >>>> Anyway, hopefully its straight forward. Also, if it is exiting >>>> cleanly, you could set a break on the return line and try and get a >>>> traceback of what lead to the exit. >>>> >>>> Also, how did you install spidermonkey exactly, was it a manual build >>>> or from yum or an rpm? >>>> >>>> Paul >>>> >>>>> On Mar 31, 2009, at 7:46 AM, Jan Lehnardt wrote: >>>>> >>>>>> Hi, >>>>>> >>>>>> On 31 Mar 2009, at 06:39, Peter Hsu wrote: >>>>>> >>>>>>> Ran your suggestion.. >>>>>>> >>>>>>> [r...@rhel4-dell 0.9.0]# su - couchdb >>>>>>> [couc...@rhel4-dell ~]$ curl >>>>>>> http://127.0.0.1:5984/_config/query_servers >>>>>>> {"javascript":"/usr/local/bin/couchjs /usr/local/share/couchdb/ >>>>>>> server/main.js"} >>>>>>> [couc...@rhel4-dell ~]$ /usr/local/bin/couchjs /usr/local/share/ >>>>>>> couchdb/server/main.js >>>>>>> ["reset"] >>>>>>> true >>>>>>> >>>>>>> The test suite still returns the same errors. >>>>>> >>>>>> Maybe your LD_LIBRARY PATH is hosed for the CouchDB user and the >>>>>> couchjs process >>>>>> can't find the libs (libjs.so, libcurl.so) it links? >>>>>> >>>>>> Cheers >>>>>> Jan >>>>>> -- >>>>>> >>>>>> >>>>>> >>>>>>> >>>>>>> On Mar 30, 2009, at 8:49 PM, Chris Anderson wrote: >>>>>>> >>>>>>>> On Mon, Mar 30, 2009 at 7:33 PM, Vinubalaji Gopal >>>>>>>> <[email protected] >>>>>>>>> >>>>>>>>> wrote: >>>>>>>>> Hi all, >>>>>>>>> I have been trying to run couch 0.9 on RHEL 4 and I am getting >>>>>>>>> errors when trying to run the test suite. I installed spider monkey >>>>>>>>> 1.7.0 and also a newer version of curl since couchdb configure >>>>>>>>> script >>>>>>>>> complained that I needed curl >= 7.15.5. I tried almost >>>>>>>>> everything I >>>>>>>>> could think of - but I can't get rid of these errors! >>>>>>>>> >>>>>>>> >>>>>>>> I don't have the ability to reproduce this setup currently, but the >>>>>>>> error message you see looks like the couchjs process isn't launching >>>>>>>> >>>>>>>> Try running this command: >>>>>>>> >>>>>>>> curl http://127.0.0.1:5984/_config/query_servers >>>>>>>> >>>>>>>> And then run the command quoted in the output. When it's working >>>>>>>> properly, you should be able to type: >>>>>>>> >>>>>>>> ["reset"] >>>>>>>> >>>>>>>> and get the response: >>>>>>>> >>>>>>>> true >>>>>>>> >>>>>>>>> Here is the output when I run the basic test in the test suite: >>>>>>>>> Exception raised: >>>>>>>>> {"error":"{bad_return_value,{os_process_error,{exit_status, >>>>>>>>> 127}}}","reason":"{gen_server,call,[couch_query_servers, >>>>>>>>> {get_proc,<<\"javascript\">>}]}"} >>>>>>>>> >>>>>>>>> Reduce fails with the following exception: >>>>>>>>> Exception raised: >>>>>>>>> {"error":"badmatch","reason":"{{bad_return_value,{os_process_error, >>>>>>>>> {exit_status,127}}},\n >>>>>>>>> {gen_server,call,[couch_query_servers,{get_proc,<<\"javascript >>>>>>>>> \">>}]}}"} >>>>>>>>> >>>>>>>>> and a lot of tests are failing. >>>>>>>>> >>>>>>>>> Here are the last few lines from the couch.log file. >>>>>>>>> >>>>>>>>> >>>>>>>>> [Tue, 31 Mar 2009 02:26:44 GMT] [error] [<0.1706.0>] ** Generic >>>>>>>>> server >>>>>>>>> <0.1706.0> terminating >>>>>>>>> ** Last message in was {'EXIT',<0.1704.0>,killed} >>>>>>>>> ** When Server state == {file_descriptor,prim_file,{#Port<0.1112>, >>>>>>>>> 19}} >>>>>>>>> ** Reason for termination == >>>>>>>>> ** killed >>>>>>>>> >>>>>>>>> >>>>>>>>> [Tue, 31 Mar 2009 02:26:44 GMT] [error] [<0.36.0>] >>>>>>>>> {error_report,<0.22.0>, >>>>>>>>> {<0.36.0>,supervisor_report, >>>>>>>>> [{supervisor,{local,couch_server_sup}}, >>>>>>>>> {errorContext,child_terminated}, >>>>>>>>> {reason,shutdown}, >>>>>>>>> {offender, >>>>>>>>> [{pid,<0.44.0>}, >>>>>>>>> {name,couch_secondary_services}, >>>>>>>>> {mfa,{couch_server_sup,start_secondary_services,[]}}, >>>>>>>>> {restart_type,permanent}, >>>>>>>>> {shutdown,infinity}, >>>>>>>>> {child_type,supervisor}]}]}} >>>>>>>>> >>>>>>>>> >>>>>>>>> [Tue, 31 Mar 2009 02:26:44 GMT] [error] [<0.1706.0>] >>>>>>>>> {error_report,<0.22.0>, >>>>>>>>> {<0.1706.0>,crash_report, >>>>>>>>> [[{pid,<0.1706.0>}, >>>>>>>>> {registered_name,[]}, >>>>>>>>> {error_info,{exit,killed, >>>>>>>>> [{gen_server,terminate,6}, >>>>>>>>> {proc_lib,init_p_do_apply,3}]}}, >>>>>>>>> {initial_call,{couch_file,init,['Argument__1']}}, >>>>>>>>> {ancestors, >>>>>>>>> [<0.1704.0>,couch_view,couch_secondary_services, >>>>>>>>> couch_server_sup,<0.1.0>]}, >>>>>>>>> {messages,[]}, >>>>>>>>> {links,[#Port<0.1112>]}, >>>>>>>>> {dictionary,[]}, >>>>>>>>> {trap_exit,true}, >>>>>>>>> {status,running}, >>>>>>>>> {heap_size,987}, >>>>>>>>> {stack_size,23}, >>>>>>>>> {reductions,1472}], >>>>>>>>> []]}} >>>>>>>>> >>>>>>>>> Please point me to the right direction and let me know if you need >>>>>>>>> any >>>>>>>>> other information. >>>>>>>>> >>>>>>>>> >>>>>>>>> Thank you. >>>>>>>>> >>>>>>>>> -- >>>>>>>>> Vinu >>>>>>>>> >>>>>>>>> In a world without fences who needs Gates? >>>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> -- >>>>>>>> Chris Anderson >>>>>>>> http://jchrisa.net >>>>>>>> http://couch.io >>>>>>> >>>>> >>> >
