Hi all,<br><br>running 8.3.0 on CentOS 5.2 (2.6.18-92.1.22.el5xen) I am "sometimes" experiencing Out of memory error on DRBD. Sometimes here really means that it is not reproducible and it is more surprising then expected occurrence. Problem is that once it happens, there is no real way to resolve it, just reboot the machine.<br>
<br>Trying to search on this issue I have found following Ubuntu link that includes flaws check of DRBD:<br><a href="http://people.ubuntu.com/~kees/bogosec/intrepid-main/drbd8/flawfinder.log">http://people.ubuntu.com/~kees/bogosec/intrepid-main/drbd8/flawfinder.log</a><br>
<pre>There may be other security vulnerabilities; review your code!<br>Flawfinder version 1.27, (C) 2001-2004 David A. Wheeler.<br>Number of dangerous functions in C/C++ ruleset: 160<br>/tmp/bogosec.temp_target.X0yWUA/src/drbd/drbd_bitmap.c:385: [2] (buffer) memcpy: Does not check for buffer overflows when copying to destination. Make sure destination can always hold the source data. <br>
/tmp/bogosec.temp_target.X0yWUA/src/drbd/drbd_bitmap.c:638: [2] (buffer) char: Statically-sized arrays can be overflowed. Perform bounds checking, use functions that limit length, or ensure that the size is larger than the maximum possible length. <br>
</pre>Is anybody aware if I can do anything to prevent this Out of memory states? top still shows enough of free memory (more than 125MB).<br>Would there possibly be some fix in 8.3.2 already? <br>Or, can I somehow destroy such device? drbd commands are blocked in such situation. Killing of processes does not help and shutdown of DRBD network interface neither.<br>
<br>Syslog output:<br>Jul 7 13:11:30 vsp9 kernel: drbd4: disk( Diskless -> Attaching ) <br>Jul 7 13:11:30 vsp9 kernel: drbd4: Starting worker thread (from cqueue/1 [168])<br>Jul 7 13:11:30 vsp9 kernel: drbd4: ASSERT( mdev->state.disk == Diskless && mdev->state.conn == StandAlone ) in /root/rpmbuild/BUILD/drbd-8.3.0/_kmod_build_xen/drbd/drbd_worker.c:1505<br>
Jul 7 13:11:30 vsp9 kernel: drbd4: drbd_bm_resize called with capacity == 0<br>Jul 7 13:11:30 vsp9 kernel: drbd4: worker terminated<br>Jul 7 13:11:30 vsp9 kernel: drbd4: Terminating worker thread<br>Jul 7 13:11:30 vsp9 kernel: drbd4: No usable activity log found.<br>
Jul 7 13:11:30 vsp9 kernel: drbd4: Method to ensure write ordering: barrier<br>Jul 7 13:11:30 vsp9 kernel: drbd4: max_segment_size ( = BIO size ) = 32768<br>Jul 7 13:11:30 vsp9 kernel: drbd4: drbd_bm_resize: (!b) in /root/rpmbuild/BUILD/drbd-8.3.0/_kmod_build_xen/drbd/drbd_bitmap.c:521<br>
Jul 7 13:11:30 vsp9 kernel: drbd4: drbd_bm_capacity: (!mdev->bitmap) in /root/rpmbuild/BUILD/drbd-8.3.0/_kmod_build_xen/drbd/drbd_bitmap.c:339<br>Jul 7 13:11:30 vsp9 kernel: drbd4: OUT OF MEMORY! Could not allocate bitmap! <6>drbd4: size = 0 KB (0 KB)<br>
Jul 7 13:11:30 vsp9 kernel: drbd4: disk( Attaching -> Diskless ) <br>Jul 7 13:11:30 vsp9 kernel: drbd4: conn( StandAlone -> Unconnected ) <br>Jul 7 13:11:30 vsp9 kernel: drbd4: Starting worker thread (from cqueue/4 [171])<br>
Jul 7 13:11:30 vsp9 kernel: drbd4: Starting receiver thread (from drbd4_worker [11421])<br>Jul 7 13:11:30 vsp9 kernel: drbd4: ASSERT( mdev->state.disk == Diskless && mdev->state.conn == StandAlone ) in /root/rpmbuild/BUILD/drbd-8.3.0/_kmod_build_xen/drbd/drbd_worker.c:1505<br>
Jul 7 13:11:30 vsp9 kernel: drbd4: receiver (re)started<br>Jul 7 13:11:30 vsp9 kernel: drbd4: conn( Unconnected -> WFConnection ) <br>Jul 7 13:11:30 vsp9 kernel: drbd4: drbd_bm_total_weight: (!b->bm_pages) in /root/rpmbuild/BUILD/drbd-8.3.0/_kmod_build_xen/drbd/drbd_bitmap.c:638<br>
Jul 7 13:12:01 vsp9 last message repeated 31 times<br>Jul 7 13:13:02 vsp9 last message repeated 61 times<br>...<br>Jul 7 14:18:38 vsp9 kernel: drbd4: aftr_isp( 1 -> 0 ) <br>Jul 7 14:18:38 vsp9 kernel: drbd4: drbd_bm_total_weight: (!b->bm_pages) in /root/rpmbuild/BUILD/drbd-8.3.0/_kmod_build_xen/drbd/drbd_bitmap.c:638<br>
Jul 7 14:19:09 vsp9 last message repeated 33 times<br>Jul 7 14:19:35 vsp9 last message repeated 26 times<br><br><br>Thanks,<br><br>Tino<br>