kern/177536: [zfs] zfs livelock (deadlock) with high write-to-disk load
Martin Birgmeier
Martin.Birgmeier at aon.at
Thu Aug 1 05:30:03 UTC 2013
The following reply was made to PR kern/177536; it has been noted by GNATS.
From: Martin Birgmeier <Martin.Birgmeier at aon.at>
To: "Christopher D. Harrison" <harrison at biostat.wisc.edu>
Cc: bug-followup at FreeBSD.org
Subject: Re: kern/177536: [zfs] zfs livelock (deadlock) with high write-to-disk
load
Date: Thu, 01 Aug 2013 07:20:50 +0200
It is raidz2, using gpt partitions each covering half of six 2 TB disks.
The remainder of these disks are mostly unused, with one containing a
small UFS root for booting, and the others some small partitions which
are used for various VirtualBox instances.
Regards,
Martin
On 07/31/13 22:32, Christopher D. Harrison wrote:
> what raidz level are you using: raidz2, raid0, raid1
> also what are the pool config (see below, my pool is called z):
>
> MOS Configuration:
> version: 28
> name: 'z'
> state: 0
> txg: 214028175
> pool_guid: 5334451366918939808
> hostid: 4266313884
> hostname: 'pisces.biostat.wisc.edu'
> hole_array[0]: 3
> vdev_children: 4
> vdev_tree:
> type: 'root'
> id: 0
> guid: 5334451366918939808
> children[0]:
> type: 'raidz'
> id: 0
> guid: 10192098197416954465
> nparity: 2
> metaslab_array: 34
> metaslab_shift: 37
> ashift: 9
> asize: 16003083796480
> is_log: 0
> create_txg: 4
> children[0]:
> type: 'disk'
> id: 0
> guid: 5442684633350386182
> path: '/dev/da3p1'
> phys_path: '/dev/da3p1'
> whole_disk: 1
> DTL: 2318341
> create_txg: 4
> children[1]:
> type: 'disk'
> id: 1
> guid: 4184769488943604229
> path: '/dev/da2p1'
> phys_path: '/dev/da2p1'
> whole_disk: 1
> DTL: 2318340
> create_txg: 4
> children[2]:
> type: 'disk'
> id: 2
> guid: 6673578809124996458
> path: '/dev/da1p1'
> phys_path: '/dev/da1p1'
> whole_disk: 1
> DTL: 2318358
> create_txg: 4
> children[3]:
> type: 'disk'
> id: 3
> guid: 14565784372994613264
> path: '/dev/da0p1'
> phys_path: '/dev/da0p1'
> whole_disk: 1
> DTL: 2318357
> create_txg: 4
> children[4]:
> type: 'disk'
> id: 4
> guid: 17127372890035360647
> path: '/dev/da6p1'
> phys_path: '/dev/da6p1'
> whole_disk: 1
> DTL: 2318356
> create_txg: 4
> children[5]:
> type: 'disk'
> id: 5
> guid: 5667576937780231702
> path: '/dev/label/disk05'
> phys_path: '/dev/label/disk05'
> whole_disk: 1
> DTL: 26389621
> create_txg: 4
> children[6]:
> type: 'disk'
> id: 6
> guid: 16545540667878636920
> path: '/dev/da5p1'
> phys_path: '/dev/da5p1'
> whole_disk: 1
> DTL: 2318354
> create_txg: 4
> children[7]:
> type: 'disk'
> id: 7
> guid: 16180883375304336701
> path: '/dev/da4p1'
> phys_path: '/dev/da4p1'
> whole_disk: 1
> DTL: 2318353
> create_txg: 4
> children[1]:
> type: 'raidz'
> id: 1
> guid: 7000405778472969514
> nparity: 2
> metaslab_array: 32
> metaslab_shift: 37
> ashift: 9
> asize: 16003083796480
> is_log: 0
> create_txg: 4
> children[0]:
> type: 'disk'
> id: 0
> guid: 11848332905648146086
> path: '/dev/da10p1'
> phys_path: '/dev/da10p1'
> whole_disk: 1
> DTL: 2318352
> create_txg: 4
> children[1]:
> type: 'disk'
> id: 1
> guid: 17180048053682905420
> path: '/dev/da9p1'
> phys_path: '/dev/da9p1'
> whole_disk: 1
> DTL: 2318351
> create_txg: 4
> children[2]:
> type: 'disk'
> id: 2
> guid: 14648362260613207535
> path: '/dev/da8p1'
> phys_path: '/dev/da8p1'
> whole_disk: 1
> DTL: 2318350
> create_txg: 4
> children[3]:
> type: 'disk'
> id: 3
> guid: 2512073543886070959
> path: '/dev/da7p1'
> phys_path: '/dev/da7p1'
> whole_disk: 1
> DTL: 2318349
> create_txg: 4
> children[4]:
> type: 'disk'
> id: 4
> guid: 9213113982598250625
> path: '/dev/da14p1'
> phys_path: '/dev/da14p1'
> whole_disk: 1
> DTL: 2318348
> create_txg: 4
> children[5]:
> type: 'disk'
> id: 5
> guid: 14965308070554000358
> path: '/dev/da13p1'
> phys_path: '/dev/da13p1'
> whole_disk: 1
> DTL: 2318347
> create_txg: 4
> children[6]:
> type: 'disk'
> id: 6
> guid: 8812761598924675500
> path: '/dev/da12p1'
> phys_path: '/dev/da12p1'
> whole_disk: 1
> DTL: 2318346
> create_txg: 4
> children[7]:
> type: 'disk'
> id: 7
> guid: 15539718945154864246
> path: '/dev/da11p1'
> phys_path: '/dev/da11p1'
> whole_disk: 1
> DTL: 2318345
> create_txg: 4
> children[2]:
> type: 'raidz'
> id: 2
> guid: 12675549095184402399
> nparity: 2
> metaslab_array: 30
> metaslab_shift: 37
> ashift: 9
> asize: 16003083796480
> is_log: 0
> create_txg: 4
> children[0]:
> type: 'disk'
> id: 0
> guid: 9157198169193782917
> path: '/dev/da16p1'
> phys_path: '/dev/da16p1'
> whole_disk: 1
> DTL: 35652327
> create_txg: 4
> children[1]:
> type: 'disk'
> id: 1
> guid: 14494767130322696735
> path: '/dev/da17p1'
> phys_path: '/dev/da17p1'
> whole_disk: 1
> DTL: 2318365
> create_txg: 4
> children[2]:
> type: 'disk'
> id: 2
> guid: 17589735305722725561
> path: '/dev/da18p1'
> phys_path: '/dev/da18p1'
> whole_disk: 1
> DTL: 2318364
> create_txg: 4
> children[3]:
> type: 'disk'
> id: 3
> guid: 14307956723966398718
> path: '/dev/da19p1'
> phys_path: '/dev/da19p1'
> whole_disk: 1
> DTL: 2318363
> create_txg: 4
> children[4]:
> type: 'disk'
> id: 4
> guid: 15965579855708087029
> path: '/dev/da20p1'
> phys_path: '/dev/da20p1'
> whole_disk: 1
> DTL: 2318362
> create_txg: 4
> children[5]:
> type: 'disk'
> id: 5
> guid: 760237787742339099
> path: '/dev/da21p1'
> phys_path: '/dev/da21p1'
> whole_disk: 1
> DTL: 2318361
> create_txg: 4
> children[6]:
> type: 'disk'
> id: 6
> guid: 2500372020986714069
> path: '/dev/da22p1'
> phys_path: '/dev/da22p1'
> whole_disk: 1
> DTL: 2318360
> create_txg: 4
> children[7]:
> type: 'disk'
> id: 7
> guid: 4544065732274116075
> path: '/dev/label/disk24'
> phys_path: '/dev/label/disk24'
> whole_disk: 1
> DTL: 5572568
> create_txg: 4
> children[3]:
> type: 'hole'
> id: 3
> guid: 0
> metaslab_array: 0
> metaslab_shift: 17
> ashift: 0
> asize: 0
> is_log: 0
> is_hole: 1
>
> On 07/31/13 14:56, Martin Birgmeier wrote:
>> No, this is not an LSI HBA, but rather
>>
>> atapci0:<JMicron JMB361 UDMA133 controller> port
>> 0xcc00-0xcc07,0xc880-0xc883,0xc800-0xc807,0xc480-0xc483,0xc400-0xc40f
>> mem 0xfe8fe000-0xfe8fffff irq 18 at device 0.0 on pci3
>> atapci1:<AHCI SATA controller> at channel -1 on atapci0
>> atapci1: AHCI v1.00 controller with 2 3Gbps ports, PM supported
>> ata2:<ATA channel> on atapci1
>> ata3:<ATA channel> on atapci1
>> ata4:<ATA channel> at channel 0 on atapci0
>> atapci2:<ATI IXP700/800 SATA300 controller> port
>> 0xa000-0xa007,0x9000-0x9003,0x8000-0x8007,0x7000-0x7003,0x6000-0x600f
>> mem 0xfe4ffc00-0xfe4fffff irq 19 at device 17.0 on pci0
>> atapci2: AHCI v1.20 controller with 6 3Gbps ports, PM supported
>> ata5:<ATA channel> at channel 0 on atapci2
>> ata6:<ATA channel> at channel 1 on atapci2
>> ata7:<ATA channel> at channel 2 on atapci2
>> ata8:<ATA channel> at channel 3 on atapci2
>> ata9:<ATA channel> at channel 4 on atapci2
>> ata10:<ATA channel> at channel 5 on atapci2
>> ada0 at ata5 bus 0 scbus3 target 0 lun 0
>> ada1 at ata6 bus 0 scbus4 target 0 lun 0
>> ada2 at ata7 bus 0 scbus5 target 0 lun 0
>> ada3 at ata8 bus 0 scbus6 target 0 lun 0
>> ada4 at ata9 bus 0 scbus7 target 0 lun 0
>> ada5 at ata10 bus 0 scbus8 target 0 lun 0
>>
>> (This is just a simple PC motherboard - ASUS M4A89GTD PRO USB3.)
>>
>> Regards,
>>
>> Martin
>>
>
>
>
More information about the freebsd-fs
mailing list