* bcache questions
@ 2014-03-06 9:33 Sitsofe Wheeler
2014-03-09 1:50 ` Vivek Dasmohapatra
0 siblings, 1 reply; 2+ messages in thread
From: Sitsofe Wheeler @ 2014-03-06 9:33 UTC (permalink / raw
To: linux-bcache
Hi,
I have a few bcache questions:
How does clear_stats work? For example if I do the following:
dd if=/dev/zero of=/dev/bcache0 bs=512 count=1 oflag=direct
grep -r . /sys/block/bcache0/bcache/stats_five_minute/
echo 1 > /sys/block/bcache0/bcache/clear_stats
grep -r . /sys/block/bcache0/bcache/stats_five_minute/
the results of the grep are the same on 3.13.5-200.fc20.x86_64 . Is this
to be expected?
Is it possible to quickly invalidate the cache so that it is known that
it is empty? Will doing
echo d8c40aa6-27b8-420e-abe2-1e2e7b31e74a > /sys/block/bcache0/bcache/detach
echo d8c40aa6-27b8-420e-abe2-1e2e7b31e74a > /sys/block/bcache0/bcache/attach
throw away the contents of the caching device?
Is it possible to invalidate the cache without destroying /dev/bcache0 ?
How do I flush the cache to backing device as quickly as possible
regardless of the usual throttling?
Should attaching the same cache set multiple times be producing WARNINGs?
[ 2043.396173] ------------[ cut here ]------------
[ 2043.396223] WARNING: CPU: 1 PID: 814 at fs/sysfs/dir.c:486 sysfs_warn_dup+0x86/0xa0()
[ 2043.396291] sysfs: cannot create duplicate filename '/devices/LNXSYSTM:00/device:00/PNP0A03:00/device:08/VMBUS:00/vmbus_0_5/host3/target3:0:0/3:0:0:4/block/sdf/bcache/cache'
[ 2043.396399] Modules linked in: iscsi_target_mod target_core_mod iscsi_tcp libiscsi_tcp libiscsi scsi_transport_iscsi crct10dif_pclmul crc32_pclmul ghash_clmulni_intel microcode serio_raw hyperv_keyboard snd_pcsp hv_netvsc i2c_piix4 i2c_core snd_pcm snd_page_alloc snd_timer snd soundcore hv_utils hyperv_fb bcache hid_hyperv hv_storvsc crc32c_intel ata_generic pata_acpi hv_vmbus dm_multipath
[ 2043.396791] CPU: 1 PID: 814 Comm: bash Not tainted 3.13.5-200.fc20.x86_64 #1
[ 2043.396880] Hardware name: Microsoft Corporation Virtual Machine/Virtual Machine, BIOS 090006 05/23/2012
[ 2043.397006] 0000000000000009 ffff8800f3233c58 ffffffff81686fdc ffff8800f3233ca0
[ 2043.397113] ffff8800f3233c90 ffffffff8106d47d ffff8800e403f000 ffff8800e403f000
[ 2043.397218] ffff88007ec5f620 0000000000000001 0000000000000000 ffff8800f3233cf0
[ 2043.397319] Call Trace:
[ 2043.397365] [<ffffffff81686fdc>] dump_stack+0x45/0x56
[ 2043.397440] [<ffffffff8106d47d>] warn_slowpath_common+0x7d/0xa0
[ 2043.397525] [<ffffffff8106d4ec>] warn_slowpath_fmt+0x4c/0x50
[ 2043.397634] [<ffffffff812307a6>] sysfs_warn_dup+0x86/0xa0
[ 2043.397710] [<ffffffff81230800>] sysfs_add_one+0x40/0x50
[ 2043.397786] [<ffffffff81231371>] sysfs_do_create_link_sd.isra.2+0xc1/0x1f0
[ 2043.397883] [<ffffffff812314c5>] sysfs_create_link+0x25/0x50
[ 2043.397976] [<ffffffffa005cf2e>] bcache_device_link+0x9e/0xf0 [bcache]
[ 2043.398080] [<ffffffffa005e832>] bch_cached_dev_attach+0x212/0x450 [bcache]
[ 2043.398184] [<ffffffffa00643cb>] bch_cached_dev_store+0x6db/0x770 [bcache]
[ 2043.398252] [<ffffffff8122f260>] ? sysfs_write_file+0x60/0x1c0
[ 2043.398299] [<ffffffff8122f328>] sysfs_write_file+0x128/0x1c0
[ 2043.398344] [<ffffffff811b8ab4>] vfs_write+0xb4/0x1f0
[ 2043.398383] [<ffffffff811b94e9>] SyS_write+0x49/0xa0
[ 2043.398423] [<ffffffff81695fa9>] system_call_fastpath+0x16/0x1b
[ 2043.398466] ---[ end trace e55ca2066db90472 ]---
[ 2043.398502] ------------[ cut here ]------------
[ 2043.398545] WARNING: CPU: 1 PID: 814 at drivers/md/bcache/super.c:689 bcache_device_link+0xbc/0xf0 [bcache]()
[ 2043.398651] Couldn't create device <-> cache set symlinks
[ 2043.398717] Modules linked in:
[ 2043.400814] iscsi_target_mod target_core_mod iscsi_tcp libiscsi_tcp libiscsi scsi_transport_iscsi crct10dif_pclmul crc32_pclmul ghash_clmulni_intel microcode serio_raw hyperv_keyboard snd_pcsp hv_netvsc i2c_piix4 i2c_core snd_pcm snd_page_alloc snd_timer snd soundcore hv_utils hyperv_fb bcache hid_hyperv hv_storvsc crc32c_intel ata_generic pata_acpi hv_vmbus dm_multipath
[ 2043.407935] CPU: 1 PID: 814 Comm: bash Tainted: G W 3.13.5-200.fc20.x86_64 #1
[ 2043.409841] Hardware name: Microsoft Corporation Virtual Machine/Virtual Machine, BIOS 090006 05/23/2012
[ 2043.413376] 0000000000000009 ffff8800f3233d00 ffffffff81686fdc ffff8800f3233d48
[ 2043.415145] ffff8800f3233d38 ffffffff8106d47d ffff8800f073008c ffff8800f0730040
[ 2043.416965] ffff88007eec0040 ffffffffa006ac98 ffff88007eec0cd0 ffff8800f3233d98
[ 2043.418686] Call Trace:
[ 2043.420428] [<ffffffff81686fdc>] dump_stack+0x45/0x56
[ 2043.422112] [<ffffffff8106d47d>] warn_slowpath_common+0x7d/0xa0
[ 2043.423784] [<ffffffff8106d4ec>] warn_slowpath_fmt+0x4c/0x50
[ 2043.425478] [<ffffffffa005cf4c>] bcache_device_link+0xbc/0xf0 [bcache]
[ 2043.427298] [<ffffffffa005e832>] bch_cached_dev_attach+0x212/0x450 [bcache]
[ 2043.428963] [<ffffffffa00643cb>] bch_cached_dev_store+0x6db/0x770 [bcache]
[ 2043.430563] [<ffffffff8122f260>] ? sysfs_write_file+0x60/0x1c0
[ 2043.432145] [<ffffffff8122f328>] sysfs_write_file+0x128/0x1c0
[ 2043.433743] [<ffffffff811b8ab4>] vfs_write+0xb4/0x1f0
[ 2043.435335] [<ffffffff811b94e9>] SyS_write+0x49/0xa0
[ 2043.436877] [<ffffffff81695fa9>] system_call_fastpath+0x16/0x1b
[ 2043.438434] ---[ end trace e55ca2066db90473 ]---
[ 2043.439973] bcache: bch_cached_dev_attach() Caching sdf as bcache0 on set d8c40aa6-27b8-420e-abe2-1e2e7b31e74a
[ 2112.863194] bcache: bch_cached_dev_attach() Can't attach sdf: already attached
[ 2112.863256] bcache: __cached_dev_store() Can't attach d8c40aa6-27b8-420e-abe2-1e2e7b31e74a
Finally, is it possible to limit the amount of cache a particular
backing device can use?
--
Sitsofe | http://sucs.org/~sits/
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: bcache questions
2014-03-06 9:33 bcache questions Sitsofe Wheeler
@ 2014-03-09 1:50 ` Vivek Dasmohapatra
0 siblings, 0 replies; 2+ messages in thread
From: Vivek Dasmohapatra @ 2014-03-09 1:50 UTC (permalink / raw
To: Sitsofe Wheeler; +Cc: linux-bcache
> Finally, is it possible to limit the amount of cache a particular
> backing device can use?
I'm not sure if it is, but you might be able to get the same end result
by using partitions on the cache device to make separate sets: That is,
instead of:
{ cache: backing }
{ sdb : sda1 }
{ sdb : sda3 }
You could try
{ sdb1 : sda1 }
{ sdb2 : sda3 }
and so forth, partitioning sdb appropriately.
[ Disclaimer: I have not actually tried this ]
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2014-03-09 1:50 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2014-03-06 9:33 bcache questions Sitsofe Wheeler
2014-03-09 1:50 ` Vivek Dasmohapatra
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.