From: "Martin K. Petersen" <martin.petersen@oracle.com> To: Tim Walker <tim.t.walker@seagate.com> Cc: "Martin K. Petersen" <martin.petersen@oracle.com>, Damien Le Moal <Damien.LeMoal@wdc.com>, Ming Lei <ming.lei@redhat.com>, "linux-block\@vger.kernel.org" <linux-block@vger.kernel.org>, linux-scsi <linux-scsi@vger.kernel.org>, "linux-nvme\@lists.infradead.org" <linux-nvme@lists.infradead.org> Subject: Re: [LSF/MM/BPF TOPIC] NVMe HDD Date: Wed, 12 Feb 2020 23:17:00 -0500 [thread overview] Message-ID: <yq1r1yzqfyb.fsf@oracle.com> (raw) In-Reply-To: <CANo=J16cDBUDWdV7tdY33UO0UT0t-g7jRfMVTxZpePvLew7Mxg@mail.gmail.com> (Tim Walker's message of "Wed, 12 Feb 2020 22:12:48 -0500") Tim, > SAS currently supports QD256, but the general consensus is that most > customers don't run anywhere near that deep. Does it help the system > for the HD to report a limited (256) max queue depth, or is it really > up to the system to decide many commands to queue? People often artificially lower the queue depth to avoid timeouts. The default timeout is 30 seconds from an I/O is queued. However, many enterprise applications set the timeout to 3-5 seconds. Which means that with deep queues you'll quickly start seeing timeouts if a drive temporarily is having issues keeping up (media errors, excessive spare track seeks, etc.). Well-behaved devices will return QF/TSF if they have transient resource starvation or exceed internal QoS limits. QF will cause the SCSI stack to reduce the number of I/Os in flight. This allows the drive to recover from its congested state and reduces the potential of application and filesystem timeouts. > Regarding number of SQ pairs, I think HDD would function well with > only one. Some thoughts on why we would want >1: > -A priority-based SQ servicing algorithm that would permit > low-priority commands to be queued in a dedicated SQ. > -The host may want an SQ per actuator for multi-actuator devices. That's fine. I think we're just saying that the common practice of allocating very deep queues for each CPU core in the system will lead to problems since the host will inevitably be able to queue much more I/O than the drive can realistically complete. > Since NVMe doesn't guarantee command execution order, it seems the > zoned block version of an NVME HDD would need to support zone append. > Do you agree? Absolutely! -- Martin K. Petersen Oracle Linux Engineering
WARNING: multiple messages have this Message-ID (diff)
From: "Martin K. Petersen" <martin.petersen@oracle.com> To: Tim Walker <tim.t.walker@seagate.com> Cc: Damien Le Moal <Damien.LeMoal@wdc.com>, "Martin K. Petersen" <martin.petersen@oracle.com>, linux-scsi <linux-scsi@vger.kernel.org>, "linux-nvme@lists.infradead.org" <linux-nvme@lists.infradead.org>, Ming Lei <ming.lei@redhat.com>, "linux-block@vger.kernel.org" <linux-block@vger.kernel.org> Subject: Re: [LSF/MM/BPF TOPIC] NVMe HDD Date: Wed, 12 Feb 2020 23:17:00 -0500 [thread overview] Message-ID: <yq1r1yzqfyb.fsf@oracle.com> (raw) In-Reply-To: <CANo=J16cDBUDWdV7tdY33UO0UT0t-g7jRfMVTxZpePvLew7Mxg@mail.gmail.com> (Tim Walker's message of "Wed, 12 Feb 2020 22:12:48 -0500") Tim, > SAS currently supports QD256, but the general consensus is that most > customers don't run anywhere near that deep. Does it help the system > for the HD to report a limited (256) max queue depth, or is it really > up to the system to decide many commands to queue? People often artificially lower the queue depth to avoid timeouts. The default timeout is 30 seconds from an I/O is queued. However, many enterprise applications set the timeout to 3-5 seconds. Which means that with deep queues you'll quickly start seeing timeouts if a drive temporarily is having issues keeping up (media errors, excessive spare track seeks, etc.). Well-behaved devices will return QF/TSF if they have transient resource starvation or exceed internal QoS limits. QF will cause the SCSI stack to reduce the number of I/Os in flight. This allows the drive to recover from its congested state and reduces the potential of application and filesystem timeouts. > Regarding number of SQ pairs, I think HDD would function well with > only one. Some thoughts on why we would want >1: > -A priority-based SQ servicing algorithm that would permit > low-priority commands to be queued in a dedicated SQ. > -The host may want an SQ per actuator for multi-actuator devices. That's fine. I think we're just saying that the common practice of allocating very deep queues for each CPU core in the system will lead to problems since the host will inevitably be able to queue much more I/O than the drive can realistically complete. > Since NVMe doesn't guarantee command execution order, it seems the > zoned block version of an NVME HDD would need to support zone append. > Do you agree? Absolutely! -- Martin K. Petersen Oracle Linux Engineering _______________________________________________ linux-nvme mailing list linux-nvme@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-nvme
next prev parent reply other threads:[~2020-02-13 4:17 UTC|newest] Thread overview: 64+ messages / expand[flat|nested] mbox.gz Atom feed top 2020-02-10 19:20 [LSF/MM/BPF TOPIC] NVMe HDD Tim Walker 2020-02-10 19:20 ` Tim Walker 2020-02-10 20:43 ` Keith Busch 2020-02-10 20:43 ` Keith Busch 2020-02-10 22:25 ` Finn Thain 2020-02-10 22:25 ` Finn Thain 2020-02-11 12:28 ` Ming Lei 2020-02-11 12:28 ` Ming Lei 2020-02-11 19:01 ` Tim Walker 2020-02-11 19:01 ` Tim Walker 2020-02-12 1:47 ` Damien Le Moal 2020-02-12 1:47 ` Damien Le Moal 2020-02-12 22:03 ` Ming Lei 2020-02-12 22:03 ` Ming Lei 2020-02-13 2:40 ` Damien Le Moal 2020-02-13 2:40 ` Damien Le Moal 2020-02-13 7:53 ` Ming Lei 2020-02-13 7:53 ` Ming Lei 2020-02-13 8:24 ` Damien Le Moal 2020-02-13 8:24 ` Damien Le Moal 2020-02-13 8:34 ` Ming Lei 2020-02-13 8:34 ` Ming Lei 2020-02-13 16:30 ` Keith Busch 2020-02-13 16:30 ` Keith Busch 2020-02-14 0:40 ` Ming Lei 2020-02-14 0:40 ` Ming Lei 2020-02-13 3:02 ` Martin K. Petersen 2020-02-13 3:02 ` Martin K. Petersen 2020-02-13 3:12 ` Tim Walker 2020-02-13 3:12 ` Tim Walker 2020-02-13 4:17 ` Martin K. Petersen [this message] 2020-02-13 4:17 ` Martin K. Petersen 2020-02-14 7:32 ` Hannes Reinecke 2020-02-14 7:32 ` Hannes Reinecke 2020-02-14 14:40 ` Keith Busch 2020-02-14 14:40 ` Keith Busch 2020-02-14 16:04 ` Hannes Reinecke 2020-02-14 16:04 ` Hannes Reinecke 2020-02-14 17:05 ` Keith Busch 2020-02-14 17:05 ` Keith Busch 2020-02-18 15:54 ` Tim Walker 2020-02-18 15:54 ` Tim Walker 2020-02-18 17:41 ` Keith Busch 2020-02-18 17:41 ` Keith Busch 2020-02-18 17:52 ` James Smart 2020-02-18 17:52 ` James Smart 2020-02-19 1:31 ` Ming Lei 2020-02-19 1:31 ` Ming Lei 2020-02-19 1:53 ` Damien Le Moal 2020-02-19 1:53 ` Damien Le Moal 2020-02-19 2:15 ` Ming Lei 2020-02-19 2:15 ` Ming Lei 2020-02-19 2:32 ` Damien Le Moal 2020-02-19 2:32 ` Damien Le Moal 2020-02-19 2:56 ` Tim Walker 2020-02-19 2:56 ` Tim Walker 2020-02-19 16:28 ` Tim Walker 2020-02-19 16:28 ` Tim Walker 2020-02-19 20:50 ` Keith Busch 2020-02-19 20:50 ` Keith Busch 2020-02-14 0:35 ` Ming Lei 2020-02-14 0:35 ` Ming Lei 2020-02-12 21:52 ` Ming Lei 2020-02-12 21:52 ` Ming Lei
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=yq1r1yzqfyb.fsf@oracle.com \ --to=martin.petersen@oracle.com \ --cc=Damien.LeMoal@wdc.com \ --cc=linux-block@vger.kernel.org \ --cc=linux-nvme@lists.infradead.org \ --cc=linux-scsi@vger.kernel.org \ --cc=ming.lei@redhat.com \ --cc=tim.t.walker@seagate.com \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
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.