[DRBD-user] Re: [DRBD-announce] drbd-0.7.12.tar.gz

Ward Horner whorner98 at yahoo.com
Thu Aug 25 16:12:30 CEST 2005

Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.


Thanks so much for all your efforts with this terrific application!

I have a request for your next release. In drbd.conf you require the 
addresses of each of the nodes to be actual physical IP addresses. In 
drbd-6, I was able to use hostnames (which were ultimately defined in 
/etc/hosts), but I cannot seem to do that anymore with drbd-7. Now if I 
need to change the IP address of one of the nodes, I must edit 4 files. I 
must edit both /etc/hosts and /etc/drbd.conf on both nodes and I must edit 
multiple instances of the IP address in drbd.conf  (one for each resource). 
Previously, I just changed /etc/hosts on both nodes, restarted, and I was done.

Thanks for your consideration,


At 09:56 AM 8/25/2005, you wrote:
>Yet another DRBD release...
>0.7.12 (api:77/proto:74)
>  * The /proc/drbd code used a fixed single page buffer. This breaks
>    at about 30 configured devices. Changed that to use the seq_file
>    interface.
>  * Bind the source of TCP connections to the IP address that is
>    mentioned in the configuration.
>  * Fixed a connection flip-flop bug when the two peers used different
>    user provided sizes.
>  * Various fixes in the online-resize code path, including online shrinking
>    and handling the case that the memory allocation for the new bitmap
>    failed.
>  * Random improvements to drbdadm:
>    * The "size" option is now allowed in the disk section
>    * A new "disable-ip-verification" option for the global section
>    * The "disable-io-hints" option is not longer available.
>    * Allow "drbdadm -- --size=XXX resize r0".
>  * Fixed a potential very unlikely race condition that in the end would
>    trigger an ERR in drbd_actlog.c:607. Actually I never saw this trigger.
>  * Fixed a logic bug in _drbd_process_ee() that, paired with a race condition
>    could trigger a "ASSERT(b->n_req == set_size)" upon disconnect.
>  * Removed the "disable_io_hints" module parameter.
>  * Added the "disable_bd_claim" module parameter, to allow users, WHO
>    KNOW WHAT THEY DO, to read-access the data on the secondary node.
>  * Allow "drbdadm invalidate" only in StandAlone and Connected states.
>  * DRBD no longer goes into StandAlone mode, if there is an random network
>    error during the handshake phase, instead it retries the to connect.
>    Only critical problems, e.g. incompatible protocols will cause it
>    to go into StandAlone mode.
>This time I released the hard grip a bit and let some improvements slip
>in, that are save to apply in the stable series.
>The few bug-fixes are really non critical. Update as your time and
>schedule permits, no need to hurry.
>On the other hand, if you are up to use the online resizing feature, it
>might be better to upgrade before.
>: Dipl-Ing Philipp Reisner                      Tel +43-1-8178292-50 :
>: LINBIT Information Technologies GmbH          Fax +43-1-8178292-82 :
>: Schönbrunnerstr 244, 1120 Vienna, Austria    http://www.linbit.com :
>drbd-announce mailing list
>drbd-announce at lists.linbit.com

More information about the drbd-user mailing list