Crash when starting service related to log directory set in conf

[p-e-s-t-s]

I went from running rethinkdb on the command line to starting it as a service. I used the default conf except for changing the data and log locations. This crash happened after trying to (re)start the service.

The problem ended up being the log config change. I commented it back out to let rethinkdb use its default and then it started up just fine.

Probably an issue with permissions but @mglukhovsky asked me log the crash here.

cat /proc/version
Linux version 3.13.0-48-generic (buildd@orlo) (gcc version 4.8.2 (Ubuntu 4.8.2-19ubuntu1) ) #80-Ubuntu SMP Thu Mar 12 11:16:15 UTC 2015

rethinkdb -v
rethinkdb 2.0.4~0trusty (GCC 4.8.2)

sudo /etc/init.d/rethinkdb start
rethinkdb: instance1: The instance is not running (there is no pid file)
rethinkdb: instance1: will only listen on local network interfaces.
rethinkdb: instance1: To expose rethinkdb on the network, add the 'bind=all' option to /etc/rethinkdb/instances.d/instance1.conf
rethinkdb: instance1: Starting instance. (logging to `/var/log/rethinkdb')
Recursively removing directory /var/lib/rethinkdb/default
Previous message may not have been written to the log file (cannot open or find log file).
In recursion: removing file /var/lib/rethinkdb/default
Previous message may not have been written to the log file (cannot open or find log file).
Version: rethinkdb 2.0.4~0trusty (GCC 4.8.2)
error: Error in src/utils.cc at line 525:
Previous message may not have been written to the log file (cannot open or find log file).
error: Guarantee failed: [res == 0]  (errno 13 - Permission denied) Fatal error: failed to delete '/var/lib/rethinkdb/default'.
Previous message may not have been written to the log file (cannot open or find log file).
error: Backtrace:
Previous message may not have been written to the log file (cannot open or find log file).
error: Fri Aug  7 23:56:15 2015

       1: backtrace_t::backtrace_t() at ??:?
       2: format_backtrace(bool) at ??:?
       3: report_fatal_error(char const*, int, char const*, ...) at ??:?
       4: remove_directory_helper(char const*, stat const*, int, FTW*) at ??:?
       5: /lib/x86_64-linux-gnu/libc.so.6(+0xedd52) [0x7ff7411bcd52] at 0x7ff7411bcd52 (/lib/x86_64-linux-gnu/libc.so.6)
       6: /lib/x86_64-linux-gnu/libc.so.6(+0xee30b) [0x7ff7411bd30b] at 0x7ff7411bd30b (/lib/x86_64-linux-gnu/libc.so.6)
       7: remove_directory_recursive(char const*) at ??:?
       8: directory_lock_t::~directory_lock_t() at ??:?
       9: main_rethinkdb_porcelain(int, char**) at ??:?
       10: main+0x1f4 at ??:?
       11: __libc_start_main+0xf5 at 0x7ff7410f0ec5 (/lib/x86_64-linux-gnu/libc.so.6)
       12: /usr/bin/rethinkdb() [0x6f9249] at 0x6f9249 ()

[danielmewes]

Thanks for opening an issue @p-e-s-t-s .
This should be a fail_due_to_user_error, not a guarantee. That will avoid the ugly backtrace.

[p-e-s-t-s]

Thanks!

Just realized my output above shows I had the log set to /var/lib/rethinkdb - user error but still a crash.