[DRBD-user] drbd_md_sync_page_io(, 262144s, READ) out of range md access!

Lars Ellenberg lars.ellenberg at linbit.com
Fri Oct 31 16:22:41 CET 2014

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


On Wed, Oct 29, 2014 at 12:27:24PM +0000, James Dingwall wrote:
> Hi,
> 
> We are using drbd to replicate vm disk images.  Each virtual disk is
> configured on a separate volume (LVM) with an shared external
> metadata volume.  Recently we have changed from Ubuntu Precise
> running a 3.8 kernel to Trusty with 3.13 and now during system boot
> we see:
> 
> [4439]:drbd_md_sync_page_io(,262144s,READ) out of range md access!
> 
> However that only shows for devices which are using an index in the
> metadata device != 0.

Which you should not do anyways.
Using those does not make sense at all nowadays,
even less if you live on top of LVM anyways.

> If the sector size is 512bytes and metadata
> slots are 128Mb then this read should be from the 0 position for
> that metadata chunk suggesting that this is getting printed in
> error?  The relevant block from the kernel source in v3.13 is
> 
>         if (sector < drbd_md_first_sector(bdev) ||
>             sector + 7 > drbd_md_last_sector(bdev))
>                 dev_alert(DEV, "%s [%d]:%s(,%llus,%s) out of range
> md access!\n",
>                      current->comm, current->pid, __func__,
>                      (unsigned long long)sector, (rw & WRITE) ?
> "WRITE" : "READ");
> 
> Is my understanding correct that this is benign or is this something
> that we should worry about?

Bug fixes go into the out-of-tree DRBD first,
then at some point into the next kernel merge window.
*very* few of them will be also queued for "stable" kernel backports.

You will usually have a better supported system,
if you use the current out-of-tree DRBD module
than if you use whatever version happened to be "in-tree"
with the kernel version you chose.

> Thanks,
> James
> 
> These are all the messages recorded in syslog while the drbd devices
> are being configured.
> 
> [  435.302219] drbd: initialized. Version: 8.4.3 (api:1/proto:86-101)
> [  435.302224] drbd: srcversion: FD2AB826D8AFD9700421E57
> [  435.302226] drbd: registered as block device major 147
> [  435.924995] d-con 3b3a-420b-b760-f1ea494f154d-1-bin: Starting
> worker thread (from drbdsetup [6921])
> [  435.940596] block drbd1000: disk( Diskless -> Attaching )
> [  435.940705] d-con 3b3a-420b-b760-f1ea494f154d-1-bin: Method to
> ensure write ordering: flush
> [  435.940707] block drbd1000: max BIO size = 4096
> [  435.940711] block drbd1000: drbd_bm_resize called with capacity
> == 5244928
> [  435.940720] block drbd1000: resync bitmap: bits=655616
> words=10244 pages=21
> [  435.940722] block drbd1000: size = 2561 MB (2622464 KB)
> [  435.952494] block drbd1000: Writing the whole bitmap, size changed
> [  435.955598] block drbd1000: bitmap WRITE of 21 pages took 0 jiffies
> [  435.968355] block drbd1000: 2561 MB (655616 bits) marked
> out-of-sync by on disk bit-map.
> [  436.061811] block drbd1000: bitmap READ of 21 pages took 0 jiffies
> [  436.061824] block drbd1000: recounting of set bits took
> additional 0 jiffies
> [  436.061826] block drbd1000: 2561 MB (655552 bits) marked
> out-of-sync by on disk bit-map.
> [  436.061830] block drbd1000: disk( Attaching -> Inconsistent )
> [  436.061832] block drbd1000: attached to UUIDs
> 0000000000000004:0000000000000000:0000000000000000:0000000000000000
> [  436.122874] d-con 3b3a-420b-b760-f1ea494f154d-1-cfg: Starting
> worker thread (from drbdsetup [7008])
> [  436.142018] block drbd1002: drbdsetup
> [7008]:drbd_md_sync_page_io(,524288s,READ) out of range md access!
> [  436.142419] block drbd1002: disk( Diskless -> Attaching )
> [  436.142529] d-con 3b3a-420b-b760-f1ea494f154d-1-cfg: Method to
> ensure write ordering: flush
> [  436.142532] block drbd1002: max BIO size = 4096
> [  436.142535] block drbd1002: drbd_bm_resize called with capacity == 264192
> [  436.142538] block drbd1002: resync bitmap: bits=33024 words=516 pages=2
> [  436.142540] block drbd1002: size = 129 MB (132096 KB)
> [  436.155712] block drbd1002: Writing the whole bitmap, size changed
> [  436.155772] block drbd1002: bitmap WRITE of 2 pages took 0 jiffies
> [  436.164093] block drbd1002: 129 MB (33024 bits) marked
> out-of-sync by on disk bit-map.
> [  436.219949] block drbd1002: bitmap READ of 2 pages took 0 jiffies
> [  436.219952] block drbd1002: recounting of set bits took
> additional 0 jiffies
> [  436.219954] block drbd1002: 129 MB (32960 bits) marked
> out-of-sync by on disk bit-map.
> [  436.219958] block drbd1002: disk( Attaching -> Inconsistent )
> [  436.219960] block drbd1002: attached to UUIDs
> 0000000000000004:0000000000000000:0000000000000000:0000000000000000
> [  436.276434] d-con 3b3a-420b-b760-f1ea494f154d-1-dat: Starting
> worker thread (from drbdsetup [7051])
> [  436.278703] block drbd1003: drbdsetup
> [7051]:drbd_md_sync_page_io(,786432s,READ) out of range md access!
> [  436.278748] block drbd1003: disk( Diskless -> Attaching )
> [  436.278853] d-con 3b3a-420b-b760-f1ea494f154d-1-dat: Method to
> ensure write ordering: flush
> [  436.278856] block drbd1003: max BIO size = 4096
> [  436.278860] block drbd1003: drbd_bm_resize called with capacity
> == 8390656
> [  436.278872] block drbd1003: resync bitmap: bits=1048832
> words=16388 pages=33
> [  436.278873] block drbd1003: size = 4097 MB (4195328 KB)
> [  436.286634] block drbd1003: Writing the whole bitmap, size changed
> [  436.286957] block drbd1003: bitmap WRITE of 33 pages took 0 jiffies
> [  436.308142] block drbd1003: 4097 MB (1048832 bits) marked
> out-of-sync by on disk bit-map.
> [  436.378622] block drbd1003: bitmap READ of 33 pages took 0 jiffies
> [  436.378658] block drbd1003: recounting of set bits took
> additional 0 jiffies
> [  436.378661] block drbd1003: 4097 MB (1048768 bits) marked
> out-of-sync by on disk bit-map.
> [  436.378666] block drbd1003: disk( Attaching -> Inconsistent )
> [  436.378670] block drbd1003: attached to UUIDs
> 0000000000000004:0000000000000000:0000000000000000:0000000000000000
> [  436.415499] d-con 3b3a-420b-b760-f1ea494f154d-1-vol: Starting
> worker thread (from drbdsetup [7057])
> [  436.419093] block drbd1001: drbdsetup
> [7057]:drbd_md_sync_page_io(,262144s,READ) out of range md access!
> [  436.419187] block drbd1001: disk( Diskless -> Attaching )
> [  436.419379] d-con 3b3a-420b-b760-f1ea494f154d-1-vol: Method to
> ensure write ordering: flush
> [  436.419383] block drbd1001: max BIO size = 4096
> [  436.419388] block drbd1001: drbd_bm_resize called with capacity
> == 6293504
> [  436.419409] block drbd1001: resync bitmap: bits=786688
> words=12292 pages=25
> [  436.419412] block drbd1001: size = 3073 MB (3146752 KB)
> [  436.429188] block drbd1001: Writing the whole bitmap, size changed
> [  436.429764] block drbd1001: bitmap WRITE of 25 pages took 0 jiffies
> [  436.441135] block drbd1001: 3073 MB (786688 bits) marked
> out-of-sync by on disk bit-map.
> [  436.484556] block drbd1001: bitmap READ of 25 pages took 0 jiffies
> [  436.484584] block drbd1001: recounting of set bits took
> additional 0 jiffies
> [  436.484587] block drbd1001: 3073 MB (786624 bits) marked
> out-of-sync by on disk bit-map.
> [  436.484592] block drbd1001: disk( Attaching -> Inconsistent )
> [  436.484596] block drbd1001: attached to UUIDs
> 0000000000000004:0000000000000000:0000000000000000:0000000000000000
> 
> _______________________________________________
> drbd-user mailing list
> drbd-user at lists.linbit.com
> http://lists.linbit.com/mailman/listinfo/drbd-user

-- 
: Lars Ellenberg
: LINBIT | Your Way to High Availability
: DRBD/HA support and consulting http://www.linbit.com

DRBD® and LINBIT® are registered trademarks of LINBIT, Austria.
__
please don't Cc me, but send to list   --   I'm subscribed



More information about the drbd-user mailing list