Red Hat Bugzilla – Bug 844695
excessive logging in self-heal daemon should be reduced
Last modified: 2012-09-12 07:05:49 EDT
Created attachment 601512 [details] log analysing script which parses the log file and prints the number of times each function has logged Description of problem: Self-heal daemon logs heavily by increasing the log file to some GBs thus consuming large disk space. It should be minimized by preventing the logs being repeated and also avoiding non important logs or printing those log message in some other log level such as DEBUG or TRACE. With the bug a script has been attached which analyzes the glusterfs log file provided as an argument and prints the functions and number of times they have been called. This is the o/p of log analyzer script run on a self-heal-daemon log file which had grown to 15-20MBs. afr_detect_self_heal_by_iatt:7353 afr_detect_self_heal_by_lookup_status:243 afr_dir_exclusive_crawl:20 afr_launch_self_heal:19993 afr_notify:15 afr_self_heal_completion_cbk:19993 afr_sh_common_lookup_resp_handler:5745 afr_sh_data_open_cbk:240 afr_sh_entry_expunge_entry_cbk:1956 afr_sh_entry_fix:1 afr_sh_metadata_post_nonblocking_inodelk_cbk:486 afr_sh_post_nonblocking_entry_cbk:3 cleanup_and_exit:2 client3_1_entrylk_cbk:3 client3_1_getxattr_cbk:3 client3_1_inodelk_cbk:243 client3_1_open_cbk:240 client3_1_unlink_cbk:4 client_ping_cbk:3 client_rpc_notify:42 client_set_lk_version_cbk:54 client_setvolume_cbk:108 _crawl_proceed:12 gf_add_cmdline_options:12 main:2 mgmt_cbk_spec:4 mgmt_getspec_cbk:5 notify:81 reincarnate:1 _remove_stale_index:9 rpc_client_ping_timer_expired:3 rpc_clnt_reconfig:24 saved_frames_unwind:6 select_server_supported_programs:54 sh_loop_driver_done:18008 socket_connect_finish:24 __socket_keepalive:2 __socket_proto_state_machine:41 __socket_rwv:18 socket_server_event_handler:2 ========= Error Functions ======== afr_notify:3 afr_self_heal_completion_cbk:243 afr_sh_common_lookup_resp_handler:5745 afr_sh_data_open_cbk:240 afr_sh_metadata_post_nonblocking_inodelk_cbk:486 afr_sh_post_nonblocking_entry_cbk:3 cleanup_and_exit:2 client3_1_entrylk_cbk:3 client3_1_getxattr_cbk:3 client3_1_inodelk_cbk:243 client3_1_open_cbk:240 client3_1_unlink_cbk:4 client_ping_cbk:3 _crawl_proceed:12 _remove_stale_index:4 saved_frames_unwind:6 socket_connect_finish:24 __socket_keepalive:2 __socket_proto_state_machine:41 __socket_rwv:18 socket_server_event_handler:2 "Error function" section lists also functions which have logged in WARNING level. Version-Release number of selected component (if applicable): How reproducible: Steps to Reproduce: 1. 2. 3. Actual results: Expected results: Additional info:
Raghavendra Bhat, Please use below in your patch, so we have better filtering done using unix scripts. ---------------- Subject: [PATCH] make output 'sort -n' friendly --- log_analyzer.sh | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/log_analyzer.sh b/log_analyzer.sh index cea10b5..7e1f483 100755 --- a/log_analyzer.sh +++ b/log_analyzer.sh @@ -27,7 +27,7 @@ function analyze () for i in $(cat /tmp/uniq_function) do count=$(grep $i /tmp/functions_file | wc -l) - echo "$i:$count" + printf "%d\t%s\n" $count $i done #grep "time of crash:" $log_file 2>/dev/null; ------------- after applying above patch, try log_analyzer.sh <logfile> | sort -n, and i feel that gives more better viewing experience.
*** This bug has been marked as a duplicate of bug 844697 ***