<html>
<head>
<style><!--
.hmmessage P
{
margin:0px;
padding:0px
}
body.hmmessage
{
font-size: 10pt;
font-family:Tahoma
}
--></style>
</head>
<body class='hmmessage'>
<BR>
Iam running managed failovers and after 7 failovers, i have dmesg print as <BR>
In a managed failovers, drbd partition is unmounted first and then "drbdadm Secondary all" is invoked.<BR>
Network is not touched until drbd transition to secondary is done.<BR>
<BR>
Any help to avoid this in managed failovers is appreciated.<BR>
<BR>
drbd0: drbd_sync_handshake:<BR>drbd0: self B3846B4B8BDF8064:B3846B4B8BDF8065:E3596DA0539760BC:0ECF8BD380756C2D<BR>drbd0: peer B3846B4B8BDF8065:B3846B4B8BDF8065:E3596DA0539760BD:0ECF8BD380756C2D<BR>drbd0: uuid_compare()=0 by rule 4<BR>drbd0: No resync, but 278 bits in bitmap!<BR><BR>
<BR>
*rule_nr = 4;<BR> if (self == peer) { /* Common power [off|failure] */<BR> int rct, dc; /* roles at crash time */<BR>
rct = (test_bit(CRASHED_PRIMARY, &mdev->flags) ? 1 : 0) +<BR> (mdev->p_uuid[UUID_FLAGS] & 2);<BR> /* lowest bit is set when we were primary,<BR> * next bit (weight 2) is set when peer was primary */<BR>
MTRACE(TraceTypeUuid, TraceLvlMetrics, DUMPI(rct); );<BR>
switch (rct) {<BR> case 0: /* !self_pri && !peer_pri */ return 0;<BR> case 1: /* self_pri && !peer_pri */ return 1;<BR> case 2: /* !self_pri && peer_pri */ return -1;<BR> case 3: /* self_pri && peer_pri */<BR> dc = test_bit(DISCARD_CONCURRENT, &mdev->flags);<BR> MTRACE(TraceTypeUuid, TraceLvlMetrics, DUMPI(dc); );<BR> return dc ? -1 : 1;<BR> }<BR> }<BR><BR>
Thanks and Regards<BR>Lak<BR><BR> <BR>
<HR id=stopSpelling>
From: putcha_laks@hotmail.com<BR>To: drbd-user@lists.linbit.com<BR>Date: Wed, 15 Sep 2010 04:35:03 +0000<BR>Subject: [DRBD-user] What is the recommended recovery action/config for rules 5 and 6<BR><BR>
<META name=Generator content="Microsoft SafeHTML">
<STYLE>
.ExternalClass .ecxhmmessage P
{padding:0px;}
.ExternalClass body.ecxhmmessage
{font-size:10pt;font-family:Tahoma;}
</STYLE>
DRBD version: 8.0.16<BR>We are forced to use this version of drbd, because of the older kernel version.<BR>Every 10 failovers i run into either rule_nr = 5 or rule_nr = 6.<BR>Its been over a month since we are trying to figure out the rootcause for this, but in vain.<BR> <BR>We have terminated all apps accessing drbd partition before calling drbdadm secondary all and umount drbd partition.<BR>With this change we are able to run about 30 failovers successfully. But we have hit rule_nr 5 after this.<BR> <BR>Can any one of you help us with recommended recovery action / config change for rules 5 and 6.<BR> <BR>*rule_nr = 5;<BR> peer = mdev->p_uuid[Bitmap] & ~((u64)1);<BR> if (self == peer)<BR> return -1;<BR> *rule_nr = 6;<BR> for (i = History_start; i <= History_end; i++) {<BR> peer = mdev->p_uuid[i] & ~((u64)1);<BR> if (self == peer)<BR> return -2;<BR> }<BR><BR>Thanks and Regards<BR>Lak.<BR><BR><BR>_______________________________________________ drbd-user mailing list drbd-user@lists.linbit.com http://lists.linbit.com/mailman/listinfo/drbd-user </body>
</html>