Latest snapshot logfetch core dumps with the following error in hobbitclient.log
2006-08-05 04:59:29 xmalloc: Out of memory! Abort - core dumped
truss logfetch...
read(3, " / v a r / a d m / m e s".., 8192) = 74 read(3, 0x00034FF4, 8192) = 0 llseek(3, 0, SEEK_CUR) = 74 close(3) = 0 open("/var/adm/messages", O_RDONLY) = 3 fstat(3, 0xFFBFFB78) = 0 lseek(3, 328208, SEEK_SET) = 328208 llseek(3, 0, SEEK_CUR) = 328208 brk(0x00038FE8) = 0 brk(0x80038FE0) = 0 brk(0xFFFEAFE8) Err#12 ENOMEM brk(0x80038FE0) = 0 brk(0x00038FE8) = 0 time() = 1154753484 brk(0x00038FE8) = 0 brk(0x0003AFE8) = 0 open("/usr/share/lib/zoneinfo/GMT", O_RDONLY) = 4 fstat64(4, 0xFFBFE838) = 0 read(4, " T Z i f\0\0\0\0\0\0\0\0".., 56) = 56 close(4) = 0 fstat64(2, 0xFFBFDAC8) = 0 2006-08-05 04:51:24write(2, " 2 0 0 6 - 0 8 - 0 5 0".., 19) = 19 write(2, " ", 1) = 1 xmalloc: Out of memory! write(2, " x m a l l o c : O u t".., 24) = 24 sigaction(SIGABRT, 0x00000000, 0xFFBFFA20) = 0 llseek(3, 0, SEEK_CUR) = 328208 sigaction(SIGABRT, 0xFFBFF8D0, 0xFFBFF970) = 0 sigfillset(0xFF36FB20) = 0 schedctl() = 0xFF382000 lwp_sigmask(SIG_SETMASK, 0x00000000, 0x00000000) = 0xFFBFFEFF [0x0000FFFF] lwp_kill(1, SIGABRT) = 0 Received signal #6, SIGABRT [default] siginfo: SIGABRT pid=18172 uid=22020 code=-1
I recompiled an older snapshot from August 1, 2006 and logfetch works fine. Perhaps the truss above matches the diff between the August 5 logfetch and August 1?
root@:/var/tmp/080106/snapshot/client> diff logfetch.c ../../../snapshot/client/logfetch.c 15c15 < static char rcsid[] = "$Id: logfetch.c,v 1.35 2006/07/19 05:42:46 henrik Exp $";
static char rcsid[] = "$Id: logfetch.c,v 1.36 2006/08/04 15:57:47 henrik Exp $"; 30a31,36 /* Some systems do not have the S_ISSOCK macro for stat() */ #ifdef SCO_SV #include <cpio.h> #define S_ISSOCK(m) (((m) & S_IFMT) == C_ISSOCK) #endif
logfetch is running on Solaris 10 where it cores with the August 4 and August 5 snapshots. Truss with older logfetch:
read(3, " / v a r / a d m / m e s".., 8192) = 38 read(3, 0x000350E4, 8192) = 0 llseek(3, 0, SEEK_CUR) = 38 close(3) = 0 open64("/var/adm/messages", O_RDONLY) = 3 fstat64(3, 0xFFBFFB68) = 0 llseek(3, 0, SEEK_SET) = 0 llseek(3, 0, SEEK_CUR) = 0 fstat64(3, 0xFFBFF9A8) = 0 brk(0x000390D8) = 0 brk(0x0003B0D8) = 0 fstat64(3, 0xFFBFF850) = 0 ioctl(3, TCGETA, 0xFFBFF934) Err#25 ENOTTY read(3, " A u g 3 0 6 : 4 0".., 8192) = 1426 read(3, 0x000379BC, 8192) = 0 llseek(3, 0, SEEK_CUR) = 1426 close(3) = 0 ioctl(1, TCGETA, 0xFFBFEE14) = 0 fstat64(1, 0xFFBFED30) = 0 [msgs:/var/adm/messages] write(1, " [ m s g s : / v a r / a".., 25) = 25
Which looks better.
~David
David Gore a écrit :
Latest snapshot logfetch core dumps with the following error in hobbitclient.log
I too get a core dump. I tried to remove SCO's defines, it still crashes (seems logical).
Howerver here's what I get :
2006-08-07 15:42:07 LFSfile /var/adm/messages: st_size 40862318854144, ofs[0] 2047, ofs[6] -60307452578564912 2006-08-07 15:42:07 Segmentation Fault - core dumped
I noticed gcc complains about non-matching types in : gcc -g -O2 -Wall -Wno-unused -D_REENTRANT -DSunOS -I. -I/usr/local/src/snapshot/include -DCLIENTONLY=1 -o logfetch logfetch.c ../lib/hobbitclient.a logfetch.c: In function `savelogstatus': logfetch.c:834: warning: long long int format, different type arg (arg 3)
(but this line didn't move for a long time)
-- Charles Goyard - cgoyard at cvf.fr - (+33) 1 45 38 01 31
participants (2)
-
cgoyard@cvf.fr
-
David.Gore@VerizonBusiness.com