|
Date: 2/16/2008 2:56 pm · Subject: Re: Database cache causes segfaults · Rating: 0
I have confirmed that this still occurs on WRE 0.8.1 / WebGUI 7.4.24. Here's the strace output for a site where the database is named "webgui_envco" and it lives on a MySQL server running on 192.168.77.164 (from the WebGUI mod_perl server's perspective): 28549 connect(19, {sa_family=AF_INET, sin_port=htons(3306), sin_addr=inet_addr("192.168.77.164")}, 16) = 0 28549 setsockopt(19, SOL_SOCKET, SO_RCVTIMEO, "\2003\341\1\0\0\0\0", 8) = 0 28549 setsockopt(19, SOL_SOCKET, SO_SNDTIMEO, "\2003\341\1\0\0\0\0", 8) = 0 28549 setsockopt(19, SOL_IP, IP_TOS, [8], 4) = 0 28549 setsockopt(19, SOL_TCP, TCP_NODELAY, [1], 4) = 0 28549 setsockopt(19, SOL_SOCKET, SO_KEEPALIVE, [1], 4) = 0 28549 read(19, "C\0\0\0\n5.0.51a-community-log\0001\2439\0u"..., 16384) = 71 28549 write(19, "O\0\0\1\217\242\2\0\0\0\0@\10\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 83) = 83 28549 read(19, "\1\0\0\2\376", 16384) = 5 28549 write(19, "\t\0\0\3WKNHLMUD\0", 13) = 13 28549 read(19, "\7\0\0\4\0\0\0\2\0\0\0", 16384) = 11 28549 poll([{fd=19, events=POLLIN|POLLPRI}], 1, 0) = 0 28549 write(19, "\21\0\0\0\3set autocommit=1", 21) = 21 28549 read(19, "\7\0\0\1\0\0\0\2\0\0\0", 16384) = 11 28549 open("/dev/urandom", O_RDONLY|O_LARGEFILE) = 20 28549 read(20, "\340tZ\373", 4) = 4 28549 close(20) = 0 28549 gettimeofday({1203192462, 792868}, NULL) = 0 28549 time(NULL) = 1203192462 28549 poll([{fd=19, events=POLLIN|POLLPRI}], 1, 0) = 0 28549 write(19, "\27\0\0\0\3select * from settings", 27) = 27 28549 read(19, "\1\0\0\1\2:\0\0\2\3def\fwebgui_envco\10setti"..., 16384) = 1448 28549 read(19, "viewLength\002309\0\0,!commercePurcha"..., 16384) = 1448 28549 read(19, "his email address exists in our "..., 16384) = 1087 28549 poll([{fd=19, events=POLLIN|POLLPRI}], 1, 0) = 0 28549 write(19, "G\0\0\0\3replace into userSession (s"..., 75) = 75 28549 read(19, "\7\0\0\1\0\1\0\2\0\1\0", 16384) = 11 28549 poll([{fd=19, events=POLLIN|POLLPRI}], 1, 0) = 0 28549 write(19, "\241\0\0\0\3update userSession set admi"..., 165) = 165 28549 read(19, "0\0\0\1\0\1\0\2\0\0\0(Rows matched: 1 Cha"..., 16384) = 52 28549 time(NULL) = 1203192462 28549 poll([{fd=19, events=POLLIN|POLLPRI}], 1, 0) = 0 28549 write(19, "\243\0\0\0\3select content from cache w"..., 167) = 167 28549 read(19, "\1\0\0\1\1:\0\0\2\3def\fwebgui_envco\5cache"..., 16384) = 1182 28549 --- SIGSEGV (Segmentation fault) @ 0 (0) --- 28549 chdir("/data/wre/prereqs") = 0 28549 rt_sigaction(SIGSEGV, {SIG_DFL}, {SIG_DFL}, 8) = 0 28549 kill(28549, SIGSEGV) = 0 28549 sigreturn() = ? (mask now []) 28549 --- SIGSEGV (Segmentation fault) @ 0 (0) --- I've seen many examples of this, and every time it happens right after an attempt to select from the cache database. The segfaults stop if I switch to the FileCache (but I can't run on that as my iowait skyrockets).
|
