Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
On Mon, Jan 07, 2008 at 08:56:56AM +0000, Ben Clewett wrote: > > > Hi Carson, > > Thanks for your reply. It is certainly odd that a write-ahead-cache > should make DRBD slower rather than faster. you definetely misunderstood something here. > The comments I picked up were from Lars Ellenberg: > > http://thread.gmane.org/gmane.linux.network.drbd/11397/focus=11401 > > "no. more concern about your write cache in the cciss controller, > and whether it lies to the kernel about having written stuff to disk > when in fact it is only in that (probably not battery backed?) write > cache..." my concerns here are about data reliability, and have nothing to do with anything "locking" or not. this comment was meant to say: if you use write-back (aka write-behin, aka write-ahead) caches, then they better be persistent (aka "battery backed"). nothing more, nothing less. -- : Lars Ellenberg http://www.linbit.com : : DRBD/HA support and consulting sales at linbit.com : : LINBIT Information Technologies GmbH Tel +43-1-8178292-0 : : Vivenotgasse 48, A-1120 Vienna/Europe Fax +43-1-8178292-82 : __ please use the "List-Reply" function of your email client.