Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
On 2003-12-31T17:46:12,
Lars Ellenberg <l.g.e at web.de> said:
> to avoid the common case where people think drbd is HA-filesharing:
> fail IO requests unless Primary
> syslogd in the default setting most often does a wall for KERN_EMERG.
> KERN_ALERT might be good enough, too.
Pretty please! Yes! ;-) This is very important. Failing IO requests on
the secondary is very helpful and will confuse users much less than the
current state.
> Is this acceptable?
Yes. But I'd only add it to 0.7 and not introduce new code into 0.6...
> - if(mdev->cstate < StandAlone || MINOR(bh->b_rdev) >= minor_count) {
> + if ( mdev->cstate < StandAlone
> + || mdev->state != Primary
> + || MINOR(bh->b_rdev) >= minor_count )
> + {
> + if (mdev->state != Primary) {
> + printk(KERN_EMERG DEVICE_NAME
> + "%d: not in Primary state, "
> + "failing IO-request [%s/%d]\n",
> + (int)(mdev-drbd_conf),
> + current->comm,current->pid);
> + }
This is good, but failing the device opening first would be even better?
Sincerely,
Lars Marowsky-Brée <lmb at suse.de>
--
High Availability & Clustering \ ever tried. ever failed. no matter.
SUSE Labs | try again. fail again. fail better.
Research & Development, SUSE LINUX AG \ -- Samuel Beckett