Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
On Friday 17 December 2004 17:23, Per Liden wrote:
> Hi,
>
> Here's a patch that enables the "disk { size ...; }" option in drbd.conf.
> I've found it highly useful since it removes the need to manually poke
> around with drbdsetup on systems setup to use a small portion of a
> partition.
>
> /Per
>
I consider your patch for inclusion, but I have a question:
> diff -ru drbd-0.7.6/user/drbdadm_main.c drbd-0.7.6-new/user/drbdadm_main.c
> --- drbd-0.7.6/user/drbdadm_main.c 2004-11-29 09:44:49.000000000 +0100
> +++ drbd-0.7.6-new/user/drbdadm_main.c 2004-12-15 17:47:18.325748752 +0100
> @@ -462,8 +462,11 @@
> argv[argc++]=res->me->disk;
> argv[argc++]=res->me->meta_disk;
> argv[argc++]=res->me->meta_index;
> - opt=res->disk_options;
> - make_options(opt);
> + opt=find_opt(res->disk_options,"size");
> + if(opt) {
> + argv[argc++]="-d";
> + ssprintf(argv[argc++],"%s",opt->value);
> + }
> argv[argc++]=0;
>
With chunk, the other disk options are left off from the command
line. -- Why do you want to change anything here ?!?!
> return m_system(argv,SLEEPS_LONG);
> @@ -490,7 +493,10 @@
> argv[argc++]=res->me->device;
> argv[argc++]="resize";
> opt=find_opt(res->disk_options,"size");
> - if(opt) ssprintf(argv[argc++],"--%s=%s",opt->name,opt->value);
> + if(opt) {
> + argv[argc++]="-d";
> + ssprintf(argv[argc++],"%s",opt->value);
> + }
> argv[argc++]=0;
>
the old code makes "--size=1234" the new code makes "-d" "1234" was
there a problem with the previous way ?
> return m_system(argv,SLEEPS_SHORT);
> diff -ru drbd-0.7.6/user/drbdadm_scanner.fl
> drbd-0.7.6-new/user/drbdadm_scanner.fl ---
> drbd-0.7.6/user/drbdadm_scanner.fl 2004-10-28 14:44:25.000000000 +0200 +++
> drbd-0.7.6-new/user/drbdadm_scanner.fl 2004-12-15 17:27:23.260426296 +0100
> @@ -227,7 +227,8 @@
>
> <DISK>{
> on-io-error do_assign(IO_ERROR); CP; return TK_ON_IO_ERROR;
> - {NDELIM} expect_error("'on-io-error'");
> + size do_assign(NUM); CP; return TK_SIZE;
> + {NDELIM} expect_error("one of 'on-io-error|size'");
> }
>
Probabely this is the only necessary chunk of this patch... as far as
I understand this...
-philipp
--
: 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 :