Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
On Wed, Nov 26, 2008 at 06:31:53PM +0100, Federico Simoncelli wrote: > On Wed, Nov 26, 2008 at 5:40 PM, Lars Ellenberg > <lars.ellenberg at linbit.com> wrote: > >> I suppose that the problem is my outdate-peer which just fences the > >> node without actually outdate it. > >> Even if this is the reason I think that a "Consistent" device should > >> not be allowed to become primary (fencing the remote node) without any > >> confirmation. > > > > the "confirmation" is supposed to come from the outdate-peer handler. > > a current Primary will refuse to be Outdated, > > the outdate would fail, > > denying the promotion of the merely Consistent one. > > So basically the problem is that my outdate-peer doesn't try to > outdate the remote peer but it just fences it. > To fix this behaviour I could modify the outdate-peer handler to check > the DRBD_RESOURCE dstate and return the exit code 6 if local resource > is not "UpToDate". > > What do you think? Comments are welcome. if the power supply is your only means of communication, you don't have many options. and this outdate and resource-level fencing buisiness gets cumbersome very quickly. whenever you think you covered "the" interessting scenario, "fixed" that special case, you likely broke a dozend other possible failure scenarios (and yes, we are deep in multiple failure land here already). if you return 6, drbd will (try to) outdate itself as a side effect. if you return anything else (not in the range of 3 to 7), say 10, you get a "outdate-peer helper broken, returned 10" log message from drbd, but drbd will stay "Consistent" and "DUnknown" for the peer, as to not make any assumptions. why don't you just set a high initial wait for connection timeout? wfc-timeout 172800; if within two days no-one came and told me that I'm outdated, and I still cannot reach the other node, I have all right to assume I'm the only survivor and allowed to become primary. or am I off track? -- : Lars Ellenberg : LINBIT | Your Way to High Availability : DRBD/HA support and consulting http://www.linbit.com DRBD® and LINBIT® are registered trademarks of LINBIT, Austria. __ please don't Cc me, but send to list -- I'm subscribed