Hi,
The error snippet you sent is not from the packages, but from a self
built module ("GIT-hash: 89a294209144b68adb3ee85a73221f964d3ee515
build by motih@CentOS-devel, 2013-06-19 20:36:06"). Are you sure the
packaged rpms behave the same?
In any case, if upstream decided to perform this change, there must be
a reason to do so. So undoing a change that upstream deliberately
performed is risky. I would ask the drbd developers about this issue.
Thanks for the info!
On Sun, Jun 23, 2013 at 02:05:27PM +0000, Moti Haimovski wrote:
> Hi,
> I've installed the drbd and-drbd-kmodel packages on CentOS 6.4.
> When tried to start the drbd service I've received the following error message:
> drbd: initialized. Version: 8.4.3 (api:1/proto:86-101)
> drbd: GIT-hash: 89a294209144b68adb3ee85a73221f964d3ee515 build by motih@CentOS-devel, 2013-06-19 20:36:06
> drbd: registered as block device major 147
> d-con r0: Starting worker thread (from drbdsetup [7617])
> block drbd0: meta data offsets don't make sense: idx=0 al_s=1, al_sz4k=8, al_offset=8, bm_offset=72, md_size_sect=262144, la_size=0, md_capaci
> ty=270336
> block drbd0: drbd_bm_resize called with capacity == 0
> d-con r0: Terminating drbd_w_r0
>
> The issue was traced to "in_core->al_stripe_size_4k != MD_32kB_SECT" condition located in
> check_offsets_and_sizes routine in drbd_main.c:
> /* old fixed size meta data is exactly that: fixed. */
> if (in_core->meta_dev_idx >= 0) {
> if (in_core->md_size_sect != MD_128MB_SECT
> || in_core->al_offset != MD_4kB_SECT
> || in_core->bm_offset != MD_4kB_SECT + MD_32kB_SECT
> || in_core->al_stripes != 1
> || in_core->al_stripe_size_4k != MD_32kB_SECT)
> goto err;
> }
>
> This parameter is initialized with MD_32kB_SECT/8 therefore the test failed,
> Looking at previous release/s - they all had the following test: "|| in_core->al_stripe_size_4k != MD_32kB_SECT/8)"
>
> I've updated my 8.4.3 code to that and it works fine now.
> block drbd0: max BIO size = 4096
> block drbd0: drbd_bm_resize called with capacity == 7538688
> block drbd0: resync bitmap: bits=942336 words=14724 pages=29
> block drbd0: size = 3681 MB (3769344 KB)
> block drbd0: Writing the whole bitmap, size changed
> block drbd0: bitmap WRITE of 29 pages took 23 jiffies
> block drbd0: 3681 MB (942336 bits) marked out-of-sync by on disk bit-map.
> block drbd0: bitmap READ of 29 pages took 0 jiffies
> block drbd0: recounting of set bits took additional 0 jiffies
> block drbd0: 3681 MB (942272 bits) marked out-of-sync by on disk bit-map.
>
> Can you approve that indeed this is the issue ?
> Are there any other issues in this release that I should be aware of ?
>
> Thanks
> Moti H.
--
Axel.Thimm at ATrpms.net
_______________________________________________
atrpms-devel mailing list
atrpms-devel@atrpms.net
http://lists.atrpms.net/mailman/listinfo/atrpms-devel
The error snippet you sent is not from the packages, but from a self
built module ("GIT-hash: 89a294209144b68adb3ee85a73221f964d3ee515
build by motih@CentOS-devel, 2013-06-19 20:36:06"). Are you sure the
packaged rpms behave the same?
In any case, if upstream decided to perform this change, there must be
a reason to do so. So undoing a change that upstream deliberately
performed is risky. I would ask the drbd developers about this issue.
Thanks for the info!
On Sun, Jun 23, 2013 at 02:05:27PM +0000, Moti Haimovski wrote:
> Hi,
> I've installed the drbd and-drbd-kmodel packages on CentOS 6.4.
> When tried to start the drbd service I've received the following error message:
> drbd: initialized. Version: 8.4.3 (api:1/proto:86-101)
> drbd: GIT-hash: 89a294209144b68adb3ee85a73221f964d3ee515 build by motih@CentOS-devel, 2013-06-19 20:36:06
> drbd: registered as block device major 147
> d-con r0: Starting worker thread (from drbdsetup [7617])
> block drbd0: meta data offsets don't make sense: idx=0 al_s=1, al_sz4k=8, al_offset=8, bm_offset=72, md_size_sect=262144, la_size=0, md_capaci
> ty=270336
> block drbd0: drbd_bm_resize called with capacity == 0
> d-con r0: Terminating drbd_w_r0
>
> The issue was traced to "in_core->al_stripe_size_4k != MD_32kB_SECT" condition located in
> check_offsets_and_sizes routine in drbd_main.c:
> /* old fixed size meta data is exactly that: fixed. */
> if (in_core->meta_dev_idx >= 0) {
> if (in_core->md_size_sect != MD_128MB_SECT
> || in_core->al_offset != MD_4kB_SECT
> || in_core->bm_offset != MD_4kB_SECT + MD_32kB_SECT
> || in_core->al_stripes != 1
> || in_core->al_stripe_size_4k != MD_32kB_SECT)
> goto err;
> }
>
> This parameter is initialized with MD_32kB_SECT/8 therefore the test failed,
> Looking at previous release/s - they all had the following test: "|| in_core->al_stripe_size_4k != MD_32kB_SECT/8)"
>
> I've updated my 8.4.3 code to that and it works fine now.
> block drbd0: max BIO size = 4096
> block drbd0: drbd_bm_resize called with capacity == 7538688
> block drbd0: resync bitmap: bits=942336 words=14724 pages=29
> block drbd0: size = 3681 MB (3769344 KB)
> block drbd0: Writing the whole bitmap, size changed
> block drbd0: bitmap WRITE of 29 pages took 23 jiffies
> block drbd0: 3681 MB (942336 bits) marked out-of-sync by on disk bit-map.
> block drbd0: bitmap READ of 29 pages took 0 jiffies
> block drbd0: recounting of set bits took additional 0 jiffies
> block drbd0: 3681 MB (942272 bits) marked out-of-sync by on disk bit-map.
>
> Can you approve that indeed this is the issue ?
> Are there any other issues in this release that I should be aware of ?
>
> Thanks
> Moti H.
--
Axel.Thimm at ATrpms.net
_______________________________________________
atrpms-devel mailing list
atrpms-devel@atrpms.net
http://lists.atrpms.net/mailman/listinfo/atrpms-devel