Re: [PATCH 0/8] block, bfq: extend bfq to support multi-actuator drives

[Arie van der Hoeven]

<Resending as plain text>

Adding Tyler Erickson who worked on the patch for concurrent positioning ranges that may be related.  Adding others at Seagate who worked on validating the patch as well.  (Muhammad, Michael, Andrew, Varun)

Thanks,
--Arie

From: Rory Chen <rory.c.chen@xxxxxxxxxxx>
Sent: Thursday, July 21, 2022 5:59 PM
To: Paolo Valente <paolo.valente@xxxxxxxxxx>
Cc: linux-block@xxxxxxxxxxxxxxx <linux-block@xxxxxxxxxxxxxxx>; linux-kernel@xxxxxxxxxxxxxxx <linux-kernel@xxxxxxxxxxxxxxx>; jack@xxxxxxx <jack@xxxxxxx>; andrea.righi@xxxxxxxxxxxxx <andrea.righi@xxxxxxxxxxxxx>; glen.valante@xxxxxxxxxx <glen.valante@xxxxxxxxxx>; Arie van der Hoeven <arie.vanderhoeven@xxxxxxxxxxx>; axboe@xxxxxxxxx <axboe@xxxxxxxxx>
Subject: RE: [PATCH 0/8] block, bfq: extend bfq to support multi-actuator drives

Hi Paolo,

I’m from Seagate China and face a problem when I’m evaluating the bfq patches. Could you please check ? Thanks

Issue statement
When running performance test on bfq patch, I observed warning message “bfq_actuator_index: bio sector out of ranges: end=35156656128” and OS hung suddenly after some hours.
The warning message is reported from function bfq_actuator_index which determines IO request is in which index of actuators.  The bio_end_sector is 35156656128 but the max LBA for the drive is 35156656127 so it’s beyond the LBA range.  I captured the block trace and didn’t found request LBA 35156656128 instead only found max request LBA 35156656127.
I’m not sure if this warning message is related to later OS hung.  /var/log/messages file is attached.


Problem environment
Kernel base is 5.18.9
Test HDD drive is Seagate ST18000NM0092 dual actuator SATA.
Actuator LBA mapping by reading VPD B9
Concurrent positioning ranges VPD page (SBC):
LBA range number: 0
number of storage elements: 1
starting LBA: 0x0
number of LBAs: 0x417c00000 [17578328064]
LBA range number: 1
number of storage elements: 1
starting LBA: 0x417c00000
number of LBAs: 0x417c00000 [17578328064]



Seagate Internal
From: Arie van der Hoeven <arie.vanderhoeven@xxxxxxxxxxx>
Sent: Wednesday, July 20, 2022 10:41 PM
To: Rory Chen <rory.c.chen@xxxxxxxxxxx>
Subject: Fw: [PATCH 0/8] block, bfq: extend bfq to support multi-actuator drives



From: Paolo Valente <paolo.valente@xxxxxxxxxx>
Sent: Thursday, June 23, 2022 8:53 AM
To: Jens Axboe <axboe@xxxxxxxxx>
Cc: linux-block@xxxxxxxxxxxxxxx <linux-block@xxxxxxxxxxxxxxx>; linux-kernel@xxxxxxxxxxxxxxx <linux-kernel@xxxxxxxxxxxxxxx>; jack@xxxxxxx <jack@xxxxxxx>; andrea.righi@xxxxxxxxxxxxx <andrea.righi@xxxxxxxxxxxxx>; glen.valante@xxxxxxxxxx <glen.valante@xxxxxxxxxx>; Arie van der Hoeven <arie.vanderhoeven@xxxxxxxxxxx>; Paolo Valente <paolo.valente@xxxxxxxxxx>
Subject: [PATCH 0/8] block, bfq: extend bfq to support multi-actuator drives


This message has originated from an External Source. Please use proper judgment and caution when opening attachments, clicking links, or responding to this email.


Hi,
this patch series extends BFQ so as to optimize I/O dispatch to
multi-actuator drives. In particular, this extension addresses the
following issue. Multi-actuator drives appear as a single device to
the I/O subsystem [1].  Yet they address commands to different
actuators internally, as a function of Logical Block Addressing
(LBAs). A given sector is reachable by only one of the actuators. For
example, Seagate’s Serial Advanced Technology Attachment (SATA)
version contains two actuators and maps the lower half of the SATA LBA
space to the lower actuator and the upper half to the upper actuator.

Evidently, to fully utilize actuators, no actuator must be left idle
or underutilized while there is pending I/O for it. To reach this
goal, the block layer must somehow control the load of each actuator
individually. This series enriches BFQ with such a per-actuator
control, as a first step. Then it also adds a simple mechanism for
guaranteeing that actuators with pending I/O are never left idle.

See [1] for a more detailed overview of the problem and of the
solutions implemented in this patch series. There you will also find
some preliminary performance results.

Thanks,
Paolo

[1] https://secure-web.cisco.com/1hcxnN1C3h1nW7mby7S66_LE8szirQwbQI0fBpYePrA0GTWfyuQyl0GpZaOn32xMSkNT0BUQWloDHFzZ23aYDZdi8NfdrEFLY9pQDBblIvn08LRiTVoIOUC8zWSG_r2PCyLtx3ppZq5cWOib_8azxteRRcbKWGdbLPSqg9hfSJSqltth0ByLONHEoI3p3e9QNIn6nVAeQbsT3aOQe-F95XrQvaPrFJXx6RGL9kDXyfkbXIHcdcLBf895gYBFn5S2WjBDQq2kzDzZOlc1HekRUhg0qDQcFY6NydVfrqNfLbpAHAth6KyREscQhVTMVREEVa1b6bQByX6grF5pn3pTIo0lODyfX6yRmcbReSYNfOZ65ZPvp-nH530FQ-5nXoRxFf74WIKDrNTALs3xQvg03DH4jLez-T2M9xEu-sfEDAEdTGF7BcnmBW6vrPO4_p3k4/https%3A%2F%2Fwww.linaro.org%2Fblog%2Fbudget-fair-queueing-bfq-linux-io-scheduler-optimizations-for-multi-actuator-sata-hard-drives%2F

Davide Zini (3):
  block, bfq: split also async bfq_queues on a per-actuator basis
  block, bfq: inject I/O to underutilized actuators
  block, bfq: balance I/O injection among underutilized actuators

Federico Gavioli (1):
  block, bfq: retrieve independent access ranges from request queue

Paolo Valente (4):
  block, bfq: split sync bfq_queues on a per-actuator basis
  block, bfq: forbid stable merging of queues associated with different
    actuators
  block, bfq: turn scalar fields into arrays in bfq_io_cq
  block, bfq: turn BFQ_NUM_ACTUATORS into BFQ_MAX_ACTUATORS

 block/bfq-cgroup.c  |  97 +++++----
 block/bfq-iosched.c | 488 +++++++++++++++++++++++++++++---------------
 block/bfq-iosched.h | 149 ++++++++++----
 block/bfq-wf2q.c    |   2 +-
 4 files changed, 493 insertions(+), 243 deletions(-)

--
2.20.1

Seagate Internal

Seagate Internal