Hide Forgot
saw this crash backtrace in a log file. 2009-06-19 16:02:06 E [posix.c:2663:posix_xattrop] posix-sdl1: /vol639/vol1298/shots/fan/char.peeps.tex/pix/ptex/arb_hi/color/v 1/512sq_dt8_tif: Bad address pending frames: frame : type(1) op(STAT) patchset: 7b2e459db65edd302aa12476bc73b3b7a17b1410 signal received: 6 configuration details:argp 1 backtrace 1 dlfcn 1 fdatasync 1 libpthread 1 llistxattr 1 setfsid 1 spinlock 1 epoll.h 1 xattr.h 1 st_atim.tv_nsec 1 package-string: glusterfs 2.0.0 /lib64/libc.so.6[0x3595e30280] /lib64/libc.so.6(gsignal+0x35)[0x3595e30215] /lib64/libc.so.6(abort+0x110)[0x3595e31cc0] /lib64/libc.so.6[0x3595e6a7fb] /lib64/libc.so.6[0x3595e700b3] /lib64/libc.so.6[0x3595e723fd] /lib64/libc.so.6(__libc_calloc+0xc0)[0x3595e73ce0] /opt/gluster-2.0.0/lib/glusterfs/2.0.0/xlator/cluster/replicate.so(afr_stat+0x69)[0x2b8cfce263d9] /opt/gluster-2.0.0/lib/glusterfs/2.0.0/xlator/cluster/distribute.so(dht_stat+0x13d)[0x2b8cfd055c4d] /opt/gluster-2.0.0/lib/glusterfs/2.0.0/xlator/performance/write-behind.so(wb_stat+0xfe)[0x2b8cfd260e1e] /opt/gluster-2.0.0/lib/glusterfs/2.0.0/xlator/mount/fuse.so[0x2b8cfd46dffd] /lib64/libpthread.so.0[0x3596606367] /lib64/libc.so.6(clone+0x6d)[0x3595ed2f7d] one of the subvolume of the afr is remote, other is local (with io-threads).
seems to be similar bug which was reported in bug id 29 and also a bug to which fix was committed in 49523c834f075007352d07bf09463af9499b59f5. I'll be sending a fix to this by end of day and we can check whether that fixes the issue. Gowda has already sent a patch which uses copies of dictionary during stack_wind of an fop to each of its children, but I am holding it back to include it in a patchset which fixes the bug in all cluster translators.
Fixed by http://patches.gluster.com/patch/652/