Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
I'm using 0.7.19 on two Fedora Core machines that I've recently
upgraded the raid card for. Before, with the Ultra Crappy card (aka
Adaptec 2610SA), I couldn't sync faster than 20MB/s over gigabit, and
doing that would completely swamp my CPU with io wait time. With my
new card I'm doing my initial sync at about 63MB/s over the same
gigabit link. I would be much happier with this 3 fold improvement,
except for this problem below that showed up in the logs:
Jun 16 00:54:50 cash kernel: Pid: 4099, comm: xvd 3 93:03 Not tainted
2.6.16-1.2133_FC5xen0 #1
Jun 16 00:54:50 cash kernel: RIP: e030:[<ffffffff8017c8ff>]
<ffffffff8017c8ff>{__bio_clone+46}
Jun 16 00:54:50 cash kernel: RSP: e02b:ffff8801de4459c8 EFLAGS:
00010216
Jun 16 00:54:50 cash kernel: RAX: ffff8801b6937f50 RBX:
ffff88006699adc0 RCX: 0000000000000050
Jun 16 00:54:50 cash kernel: RDX: ffff8801b6937e80 RSI:
ffff8801b6938000 RDI: ffff880196b5d630
Jun 16 00:54:50 cash kernel: RBP: ffff8801eeef3338 R08:
ffff8801ee82f200 R09: ffff8800028b5000
Jun 16 00:54:50 cash kernel: R10: 0000000000001000 R11:
ffffffff801f59c5 R12: 0000000000003dab
Jun 16 00:54:50 cash kernel: R13: ffff8801b6937e80 R14:
ffff8801eb73a618 R15: 000000000000000b
Jun 16 00:54:50 cash kernel: FS: 00002aaaaaab51e0(0000)
GS:ffffffff804b5000(0000) knlGS:0000000000000000
Jun 16 00:54:50 cash kernel: CS: e033 DS: 0000 ES: 0000
Jun 16 00:54:50 cash kernel: Process xvd 3 93:03 (pid: 4099,
threadinfo ffff8801de444000, task ffff8801dfd5c7d0)
Jun 16 00:54:50 cash kernel: Stack: ffff8801b6937e80 ffff88006699adc0
ffff8801b6937e80 ffffffff8017c991
Jun 16 00:54:50 cash kernel: ffff8801e55b9938 0000000000003dab
ffff8801e55b9938 ffffffff881efad5
Jun 16 00:54:50 cash kernel: 0000000007b57ffa 000000018033c9ec
Jun 16 00:54:50 cash kernel: Call Trace: <ffffffff8017c991>{bio_clone
+53} <ffffffff881efad5>{:drbd:drbd_make_request_26+1043}
Jun 16 00:54:50 cash kernel: <ffffffff881ef8ae>
{:drbd:drbd_make_request_26+492} <ffffffff801ee870>
{generic_make_request+365}
Jun 16 00:54:50 cash kernel: <ffffffff801f03d0>{submit_bio
+186} <ffffffff8010728a>{hypercall_page+650}
Jun 16 00:54:50 cash kernel: <ffffffff8026ebe4>
{dispatch_rw_block_io+1040} <ffffffff801ef2a8>{__generic_unplug_device
+19}
Jun 16 00:54:50 cash kernel: <ffffffff801427ca>{prepare_to_wait
+32} <ffffffff8026ef15>{blkif_schedule+713}
Jun 16 00:54:50 cash kernel: <ffffffff80142661>
{autoremove_wake_function+0} <ffffffff8014226f>{keventd_create_kthread
+0}
Jun 16 00:54:50 cash kernel: <ffffffff8026ec4c>{blkif_schedule
+0} <ffffffff8014226f>{keventd_create_kthread+0}
Jun 16 00:54:50 cash kernel: <ffffffff8014252a>{kthread+212}
<ffffffff8010b7f2>{child_rip+8}
Jun 16 00:54:50 cash kernel: <ffffffff8014226f>
{keventd_create_kthread+0} <ffffffff80142456>{kthread+0}
Jun 16 00:54:50 cash kernel: <ffffffff8010b7ea>{child_rip+0}
Jun 16 00:54:50 cash kernel:
Jun 16 00:54:50 cash kernel: Code: f3 a4 48 8b 02 48 89 03 48 8b 42
10 48 89 43 10 48 83 4b 18
Jun 16 00:54:50 cash kernel: RIP <ffffffff8017c8ff>{__bio_clone+46}
RSP <ffff8801de4459c8>
Jun 16 00:54:50 cash kernel: CR2: ffff8801b6938000
It doesn't look like it affected anything, but it doesn't look
reassuring, either. Is anybody else seeing this?