Re: [Lsf-pc] [LSF/MM TOPIC] iSCSI MQ adoption via MCS discussion
On 1/12/2015 2:56 PM, Bart Van Assche wrote: On 01/11/15 10:40, Sagi Grimberg wrote: I would say there is no need for specific coordination from iSCSI PoV. This is exactly what flow steering is designed for. As I see it, in order to get the TX/RX to match rings, the user can attach 5-tuple rules (using standard ethtool) to steer packets to the right rings. Hello Sagi, Can the 5-tuple rules be chosen such that it is guaranteed that the sockets used to implement per-CPU queues are spread evenly over MSI-X completion vectors ? If not, would it help to add a socket option to the Linux network stack that allows to select the TX ring explicitly, just like ib_create_cq() in the Linux RDMA stack allows to select a completion vector explicitly ? My concerns are as follows: - If the number of queues exceeds the number of MSI-X vectors then I expect that it will be much easier to guarantee even spreading by selecting tx queues explicitly instead of relying on a hashing scheme. - On multi-socket systems it is important to process completion interrupts on the CPU socket from where the I/O was initiated. I'm not sure it is possible to guarantee this when using a hashing algorithm to select the TX ring. Hey Bart, Your concerns are correct. Flow steering rules will guarantee that each socket will have a different TX/RX ring, but not necessarily the correct TX/RX ring. These issues have been addressed in the Networking subsystem. Thinking more on this out loud, There is the TX challenge, getting the HW queue selection to match the TX ring selection (which might not be the same according to flow hash), First thing that comes to mind is XPS (Transmit Packet Steering). From Documentation/networking/scaling.txt: Transmit Packet Steering is a mechanism for intelligently selecting which transmit queue to use when transmitting a packet on a multi-queue device. To accomplish this, a mapping from CPU to hardware queue(s) is recorded. The goal of this mapping is usually to assign queues exclusively to a subset of CPUs, where the transmit completions for these queues are processed on a CPU within this set. About the RX challenge, I think RFS (Receive Flow Steering) will probably be the best fit here since RX packets will be steered to the CPU where the application is running. From Documentation/networking/scaling.txt: The goal of RFS is to increase datacache hitrate by steering kernel processing of packets to the CPU where the application thread consuming the packet is running. RFS relies on the same RPS mechanisms to enqueue packets onto the backlog of another CPU and to wake up that CPU. In RFS, packets are not forwarded directly by the value of their hash, but the hash is used as index into a flow lookup table. This table maps flows to the CPUs where those flows are being processed. This definitely needs some more thinking. CC'ing Or Gerlitz which has a lot of experience in the Networking stack... Sagi. -- You received this message because you are subscribed to the Google Groups open-iscsi group. To unsubscribe from this group and stop receiving emails from it, send an email to open-iscsi+unsubscr...@googlegroups.com. To post to this group, send email to open-iscsi@googlegroups.com. Visit this group at http://groups.google.com/group/open-iscsi. For more options, visit https://groups.google.com/d/optout.
Re: [Lsf-pc] [LSF/MM TOPIC] iSCSI MQ adoption via MCS discussion
On 1/12/2015 10:05 PM, Mike Christie wrote: On 01/11/2015 03:23 AM, Sagi Grimberg wrote: On 1/9/2015 8:00 PM, Michael Christie wrote: SNIP Session wide command sequence number synchronization isn't something to be removed as part of the MQ work. It's a iSCSI/iSER protocol requirement. That is, the expected + maximum sequence numbers are returned as part of every response PDU, which the initiator uses to determine when the command sequence number window is open so new non-immediate commands may be sent to the target. So, given some manner of session wide synchronization is required between different contexts for the existing single connection case to update the command sequence number and check when the window opens, it's a fallacy to claim MC/S adds some type of new initiator specific synchronization overhead vs. single connection code. I think you are assuming we are leaving the iscsi code as it is today. For the non-MCS mq session per CPU design, we would be allocating and binding the session and its resources to specific CPUs. They would only be accessed by the threads on that one CPU, so we get our serialization/synchronization from that. That is why we are saying we do not need something like atomic_t/spin_locks for the sequence number handling for this type of implementation. If we just tried to do this with the old code where the session could be accessed on multiple CPUs then you are right, we need locks/atomics like how we do in the MCS case. I don't think we will want to restrict session per CPU. There is a tradeoff question of system resources. We might want to allow a user to configure multiple HW queues but still not to use too much of the system resources. So the session locks would still be used but definitely less congested... Are you talking about specifically the session per CPU or also MCS and doing a connection per CPU? This applies to both. Based on the srp work, how bad do you think it will be to do a session/connection per CPU? What are you thinking will be more common? Session per 4 CPU? 2 CPUs? 8? This is a level of degree which demonstrates why we need to let the user choose. I don't think there is a magic number here, there is a tradeoff between performance and memory footprint. There is also multipath to take into account here. We could do a mq/MCS session/connection per CPU (or group of CPS) then also one of those per transport path. We could also do a mq/MCS session/connection per transport path, then bind those to specific CPUs. Or something in between. Is it a good idea to tie iSCSI implementation in multipath? I've seen deployments where multipath was not used for HA (NIC bonding was used for that). The srp implementation allowed the user to choose the number of channels per target and the default was chosen by empirical results (Bart, please correct me if I'm wrong here). Sagi. -- You received this message because you are subscribed to the Google Groups open-iscsi group. To unsubscribe from this group and stop receiving emails from it, send an email to open-iscsi+unsubscr...@googlegroups.com. To post to this group, send email to open-iscsi@googlegroups.com. Visit this group at http://groups.google.com/group/open-iscsi. For more options, visit https://groups.google.com/d/optout.
Re: [Lsf-pc] [LSF/MM TOPIC] iSCSI MQ adoption via MCS discussion
On 01/11/2015 03:23 AM, Sagi Grimberg wrote: On 1/9/2015 8:00 PM, Michael Christie wrote: SNIP Session wide command sequence number synchronization isn't something to be removed as part of the MQ work. It's a iSCSI/iSER protocol requirement. That is, the expected + maximum sequence numbers are returned as part of every response PDU, which the initiator uses to determine when the command sequence number window is open so new non-immediate commands may be sent to the target. So, given some manner of session wide synchronization is required between different contexts for the existing single connection case to update the command sequence number and check when the window opens, it's a fallacy to claim MC/S adds some type of new initiator specific synchronization overhead vs. single connection code. I think you are assuming we are leaving the iscsi code as it is today. For the non-MCS mq session per CPU design, we would be allocating and binding the session and its resources to specific CPUs. They would only be accessed by the threads on that one CPU, so we get our serialization/synchronization from that. That is why we are saying we do not need something like atomic_t/spin_locks for the sequence number handling for this type of implementation. If we just tried to do this with the old code where the session could be accessed on multiple CPUs then you are right, we need locks/atomics like how we do in the MCS case. I don't think we will want to restrict session per CPU. There is a tradeoff question of system resources. We might want to allow a user to configure multiple HW queues but still not to use too much of the system resources. So the session locks would still be used but definitely less congested... Are you talking about specifically the session per CPU or also MCS and doing a connection per CPU? Based on the srp work, how bad do you think it will be to do a session/connection per CPU? What are you thinking will be more common? Session per 4 CPU? 2 CPUs? 8? There is also multipath to take into account here. We could do a mq/MCS session/connection per CPU (or group of CPS) then also one of those per transport path. We could also do a mq/MCS session/connection per transport path, then bind those to specific CPUs. Or something in between. -- You received this message because you are subscribed to the Google Groups open-iscsi group. To unsubscribe from this group and stop receiving emails from it, send an email to open-iscsi+unsubscr...@googlegroups.com. To post to this group, send email to open-iscsi@googlegroups.com. Visit this group at http://groups.google.com/group/open-iscsi. For more options, visit https://groups.google.com/d/optout.
Re: [Lsf-pc] [LSF/MM TOPIC] iSCSI MQ adoption via MCS discussion
On 01/11/2015 03:40 AM, Sagi Grimberg wrote: On 1/9/2015 10:19 PM, Mike Christie wrote: On 01/09/2015 12:28 PM, Hannes Reinecke wrote: On 01/09/2015 07:00 PM, Michael Christie wrote: On Jan 8, 2015, at 11:03 PM, Nicholas A. Bellinger n...@linux-iscsi.org wrote: On Thu, 2015-01-08 at 15:22 -0800, James Bottomley wrote: On Thu, 2015-01-08 at 14:57 -0800, Nicholas A. Bellinger wrote: On Thu, 2015-01-08 at 14:29 -0800, James Bottomley wrote: On Thu, 2015-01-08 at 14:16 -0800, Nicholas A. Bellinger wrote: SNIP The point is that a simple session wide counter for command sequence number assignment is significantly less overhead than all of the overhead associated with running a full multipath stack atop multiple sessions. I don't see how that's relevant to issue speed, which was the measure we were using: The layers above are just a hopper. As long as they're loaded, the MQ lower layer can issue at full speed. So as long as the multipath hopper is efficient enough to keep the queues loaded there's no speed degradation. The problem with a sequence point inside the MQ issue layer is that it can cause a stall that reduces the issue speed. so the counter sequence point causes a degraded issue speed over the multipath hopper approach above even if the multipath approach has a higher CPU overhead. Now, if the system is close to 100% cpu already, *then* the multipath overhead will try to take CPU power we don't have and cause a stall, but it's only in the flat out CPU case. Not to mention that our iSCSI/iSER initiator is already taking a session wide lock when sending outgoing PDUs, so adding a session wide counter isn't adding any additional synchronization overhead vs. what's already in place. I'll leave it up to the iSER people to decide whether they're redoing this as part of the MQ work. Session wide command sequence number synchronization isn't something to be removed as part of the MQ work. It's a iSCSI/iSER protocol requirement. That is, the expected + maximum sequence numbers are returned as part of every response PDU, which the initiator uses to determine when the command sequence number window is open so new non-immediate commands may be sent to the target. So, given some manner of session wide synchronization is required between different contexts for the existing single connection case to update the command sequence number and check when the window opens, it's a fallacy to claim MC/S adds some type of new initiator specific synchronization overhead vs. single connection code. I think you are assuming we are leaving the iscsi code as it is today. For the non-MCS mq session per CPU design, we would be allocating and binding the session and its resources to specific CPUs. They would only be accessed by the threads on that one CPU, so we get our serialization/synchronization from that. That is why we are saying we do not need something like atomic_t/spin_locks for the sequence number handling for this type of implementation. Wouldn't that need to be coordinated with the networking layer? Yes. Doesn't it do the same thing, matching TX/RX queues to CPUs? Yes. Hey Hannes, Mike, I would say there is no need for specific coordination from iSCSI PoV. This is exactly what flow steering is designed for. As I see it, in order to get the TX/RX to match rings, the user can attach 5-tuple rules (using standard ethtool) to steer packets to the right rings. Sagi. If so, wouldn't we decrease bandwidth by restricting things to one CPU? We have a session or connection per CPU though, so we end up hitting the same problem you talked about last year where one hctx (iscsi session or connection's socket or nic hw queue) could get overloaded. This is what I meant in my original mail where iscsi would rely on whatever blk/mq load balancers we end up implementing at that layer to balance requests across hctxs. I'm not sure I understand, The submission flow is CPU bound. In the current single queue model both CPU X and CPU Y will end up using a single socket. In the multi-queue solution, CPU X will go to socket X and CPU Y will go to socket Y. This is equal to what we have today (if only CPU X is active) or better (if more CPUs are active). Am I missing something? I did not take Hannes's comment as comparing what we have today vs the proposal. I thought he was referring to the problem he was talking about at LSF last year and saying there could be cases where we want to spread IO across CPUs/queues and some cases where we would want to execute on the CPU we were originally submitted on. I was just saying the iscsi layer would not control that and would rely on the blk/mq layer to handle this or tell us what to do similar to what we do for the rq_affinity setting. -- You received this message because you are subscribed to the Google Groups open-iscsi group. To unsubscribe from this group and stop
Re: [Lsf-pc] [LSF/MM TOPIC] iSCSI MQ adoption via MCS discussion
On 1/9/2015 10:19 PM, Mike Christie wrote: On 01/09/2015 12:28 PM, Hannes Reinecke wrote: On 01/09/2015 07:00 PM, Michael Christie wrote: On Jan 8, 2015, at 11:03 PM, Nicholas A. Bellinger n...@linux-iscsi.org wrote: On Thu, 2015-01-08 at 15:22 -0800, James Bottomley wrote: On Thu, 2015-01-08 at 14:57 -0800, Nicholas A. Bellinger wrote: On Thu, 2015-01-08 at 14:29 -0800, James Bottomley wrote: On Thu, 2015-01-08 at 14:16 -0800, Nicholas A. Bellinger wrote: SNIP The point is that a simple session wide counter for command sequence number assignment is significantly less overhead than all of the overhead associated with running a full multipath stack atop multiple sessions. I don't see how that's relevant to issue speed, which was the measure we were using: The layers above are just a hopper. As long as they're loaded, the MQ lower layer can issue at full speed. So as long as the multipath hopper is efficient enough to keep the queues loaded there's no speed degradation. The problem with a sequence point inside the MQ issue layer is that it can cause a stall that reduces the issue speed. so the counter sequence point causes a degraded issue speed over the multipath hopper approach above even if the multipath approach has a higher CPU overhead. Now, if the system is close to 100% cpu already, *then* the multipath overhead will try to take CPU power we don't have and cause a stall, but it's only in the flat out CPU case. Not to mention that our iSCSI/iSER initiator is already taking a session wide lock when sending outgoing PDUs, so adding a session wide counter isn't adding any additional synchronization overhead vs. what's already in place. I'll leave it up to the iSER people to decide whether they're redoing this as part of the MQ work. Session wide command sequence number synchronization isn't something to be removed as part of the MQ work. It's a iSCSI/iSER protocol requirement. That is, the expected + maximum sequence numbers are returned as part of every response PDU, which the initiator uses to determine when the command sequence number window is open so new non-immediate commands may be sent to the target. So, given some manner of session wide synchronization is required between different contexts for the existing single connection case to update the command sequence number and check when the window opens, it's a fallacy to claim MC/S adds some type of new initiator specific synchronization overhead vs. single connection code. I think you are assuming we are leaving the iscsi code as it is today. For the non-MCS mq session per CPU design, we would be allocating and binding the session and its resources to specific CPUs. They would only be accessed by the threads on that one CPU, so we get our serialization/synchronization from that. That is why we are saying we do not need something like atomic_t/spin_locks for the sequence number handling for this type of implementation. Wouldn't that need to be coordinated with the networking layer? Yes. Doesn't it do the same thing, matching TX/RX queues to CPUs? Yes. Hey Hannes, Mike, I would say there is no need for specific coordination from iSCSI PoV. This is exactly what flow steering is designed for. As I see it, in order to get the TX/RX to match rings, the user can attach 5-tuple rules (using standard ethtool) to steer packets to the right rings. Sagi. If so, wouldn't we decrease bandwidth by restricting things to one CPU? We have a session or connection per CPU though, so we end up hitting the same problem you talked about last year where one hctx (iscsi session or connection's socket or nic hw queue) could get overloaded. This is what I meant in my original mail where iscsi would rely on whatever blk/mq load balancers we end up implementing at that layer to balance requests across hctxs. I'm not sure I understand, The submission flow is CPU bound. In the current single queue model both CPU X and CPU Y will end up using a single socket. In the multi-queue solution, CPU X will go to socket X and CPU Y will go to socket Y. This is equal to what we have today (if only CPU X is active) or better (if more CPUs are active). Am I missing something? Sagi. -- You received this message because you are subscribed to the Google Groups open-iscsi group. To unsubscribe from this group and stop receiving emails from it, send an email to open-iscsi+unsubscr...@googlegroups.com. To post to this group, send email to open-iscsi@googlegroups.com. Visit this group at http://groups.google.com/group/open-iscsi. For more options, visit https://groups.google.com/d/optout.
Re: [Lsf-pc] [LSF/MM TOPIC] iSCSI MQ adoption via MCS discussion
On 1/9/2015 8:00 PM, Michael Christie wrote: SNIP Session wide command sequence number synchronization isn't something to be removed as part of the MQ work. It's a iSCSI/iSER protocol requirement. That is, the expected + maximum sequence numbers are returned as part of every response PDU, which the initiator uses to determine when the command sequence number window is open so new non-immediate commands may be sent to the target. So, given some manner of session wide synchronization is required between different contexts for the existing single connection case to update the command sequence number and check when the window opens, it's a fallacy to claim MC/S adds some type of new initiator specific synchronization overhead vs. single connection code. I think you are assuming we are leaving the iscsi code as it is today. For the non-MCS mq session per CPU design, we would be allocating and binding the session and its resources to specific CPUs. They would only be accessed by the threads on that one CPU, so we get our serialization/synchronization from that. That is why we are saying we do not need something like atomic_t/spin_locks for the sequence number handling for this type of implementation. If we just tried to do this with the old code where the session could be accessed on multiple CPUs then you are right, we need locks/atomics like how we do in the MCS case. I don't think we will want to restrict session per CPU. There is a tradeoff question of system resources. We might want to allow a user to configure multiple HW queues but still not to use too much of the system resources. So the session locks would still be used but definitely less congested... Sagi. -- You received this message because you are subscribed to the Google Groups open-iscsi group. To unsubscribe from this group and stop receiving emails from it, send an email to open-iscsi+unsubscr...@googlegroups.com. To post to this group, send email to open-iscsi@googlegroups.com. Visit this group at http://groups.google.com/group/open-iscsi. For more options, visit https://groups.google.com/d/optout.
Re: [Lsf-pc] [LSF/MM TOPIC] iSCSI MQ adoption via MCS discussion
On Fri, 2015-01-09 at 19:28 +0100, Hannes Reinecke wrote: [...] I think you are assuming we are leaving the iscsi code as it is today. For the non-MCS mq session per CPU design, we would be allocating and binding the session and its resources to specific CPUs. They would only be accessed by the threads on that one CPU, so we get our serialization/synchronization from that. That is why we are saying we do not need something like atomic_t/spin_locks for the sequence number handling for this type of implementation. Wouldn't that need to be coordinated with the networking layer? Doesn't it do the same thing, matching TX/RX queues to CPUs? If so, wouldn't we decrease bandwidth by restricting things to one CPU? So this is actually one of the fascinating questions on multi-queue. Long ago, when I worked for the NCR OS group and we were bringing up the first SMP systems, we actually found that the SCSI stack went faster when bound to a single CPU. The problem in those days was lock granularity and contention, so single CPU binding eliminated that overhead. However, nowadays with modern multi-tiered caching and huge latencies for cache line bouncing, we're approaching the point where the fineness of our lock granularity is hurting performance, so it's worth re-asking the question of whether just dumping all the lock latency by single CPU binding is a worthwhile exercise. James -- You received this message because you are subscribed to the Google Groups open-iscsi group. To unsubscribe from this group and stop receiving emails from it, send an email to open-iscsi+unsubscr...@googlegroups.com. To post to this group, send email to open-iscsi@googlegroups.com. Visit this group at http://groups.google.com/group/open-iscsi. For more options, visit https://groups.google.com/d/optout.
Re: [Lsf-pc] [LSF/MM TOPIC] iSCSI MQ adoption via MCS discussion
On Jan 8, 2015, at 11:03 PM, Nicholas A. Bellinger n...@linux-iscsi.org wrote: On Thu, 2015-01-08 at 15:22 -0800, James Bottomley wrote: On Thu, 2015-01-08 at 14:57 -0800, Nicholas A. Bellinger wrote: On Thu, 2015-01-08 at 14:29 -0800, James Bottomley wrote: On Thu, 2015-01-08 at 14:16 -0800, Nicholas A. Bellinger wrote: SNIP The point is that a simple session wide counter for command sequence number assignment is significantly less overhead than all of the overhead associated with running a full multipath stack atop multiple sessions. I don't see how that's relevant to issue speed, which was the measure we were using: The layers above are just a hopper. As long as they're loaded, the MQ lower layer can issue at full speed. So as long as the multipath hopper is efficient enough to keep the queues loaded there's no speed degradation. The problem with a sequence point inside the MQ issue layer is that it can cause a stall that reduces the issue speed. so the counter sequence point causes a degraded issue speed over the multipath hopper approach above even if the multipath approach has a higher CPU overhead. Now, if the system is close to 100% cpu already, *then* the multipath overhead will try to take CPU power we don't have and cause a stall, but it's only in the flat out CPU case. Not to mention that our iSCSI/iSER initiator is already taking a session wide lock when sending outgoing PDUs, so adding a session wide counter isn't adding any additional synchronization overhead vs. what's already in place. I'll leave it up to the iSER people to decide whether they're redoing this as part of the MQ work. Session wide command sequence number synchronization isn't something to be removed as part of the MQ work. It's a iSCSI/iSER protocol requirement. That is, the expected + maximum sequence numbers are returned as part of every response PDU, which the initiator uses to determine when the command sequence number window is open so new non-immediate commands may be sent to the target. So, given some manner of session wide synchronization is required between different contexts for the existing single connection case to update the command sequence number and check when the window opens, it's a fallacy to claim MC/S adds some type of new initiator specific synchronization overhead vs. single connection code. I think you are assuming we are leaving the iscsi code as it is today. For the non-MCS mq session per CPU design, we would be allocating and binding the session and its resources to specific CPUs. They would only be accessed by the threads on that one CPU, so we get our serialization/synchronization from that. That is why we are saying we do not need something like atomic_t/spin_locks for the sequence number handling for this type of implementation. If we just tried to do this with the old code where the session could be accessed on multiple CPUs then you are right, we need locks/atomics like how we do in the MCS case. -- You received this message because you are subscribed to the Google Groups open-iscsi group. To unsubscribe from this group and stop receiving emails from it, send an email to open-iscsi+unsubscr...@googlegroups.com. To post to this group, send email to open-iscsi@googlegroups.com. Visit this group at http://groups.google.com/group/open-iscsi. For more options, visit https://groups.google.com/d/optout.
Re: [Lsf-pc] [LSF/MM TOPIC] iSCSI MQ adoption via MCS discussion
On 01/09/2015 12:28 PM, Hannes Reinecke wrote: On 01/09/2015 07:00 PM, Michael Christie wrote: On Jan 8, 2015, at 11:03 PM, Nicholas A. Bellinger n...@linux-iscsi.org wrote: On Thu, 2015-01-08 at 15:22 -0800, James Bottomley wrote: On Thu, 2015-01-08 at 14:57 -0800, Nicholas A. Bellinger wrote: On Thu, 2015-01-08 at 14:29 -0800, James Bottomley wrote: On Thu, 2015-01-08 at 14:16 -0800, Nicholas A. Bellinger wrote: SNIP The point is that a simple session wide counter for command sequence number assignment is significantly less overhead than all of the overhead associated with running a full multipath stack atop multiple sessions. I don't see how that's relevant to issue speed, which was the measure we were using: The layers above are just a hopper. As long as they're loaded, the MQ lower layer can issue at full speed. So as long as the multipath hopper is efficient enough to keep the queues loaded there's no speed degradation. The problem with a sequence point inside the MQ issue layer is that it can cause a stall that reduces the issue speed. so the counter sequence point causes a degraded issue speed over the multipath hopper approach above even if the multipath approach has a higher CPU overhead. Now, if the system is close to 100% cpu already, *then* the multipath overhead will try to take CPU power we don't have and cause a stall, but it's only in the flat out CPU case. Not to mention that our iSCSI/iSER initiator is already taking a session wide lock when sending outgoing PDUs, so adding a session wide counter isn't adding any additional synchronization overhead vs. what's already in place. I'll leave it up to the iSER people to decide whether they're redoing this as part of the MQ work. Session wide command sequence number synchronization isn't something to be removed as part of the MQ work. It's a iSCSI/iSER protocol requirement. That is, the expected + maximum sequence numbers are returned as part of every response PDU, which the initiator uses to determine when the command sequence number window is open so new non-immediate commands may be sent to the target. So, given some manner of session wide synchronization is required between different contexts for the existing single connection case to update the command sequence number and check when the window opens, it's a fallacy to claim MC/S adds some type of new initiator specific synchronization overhead vs. single connection code. I think you are assuming we are leaving the iscsi code as it is today. For the non-MCS mq session per CPU design, we would be allocating and binding the session and its resources to specific CPUs. They would only be accessed by the threads on that one CPU, so we get our serialization/synchronization from that. That is why we are saying we do not need something like atomic_t/spin_locks for the sequence number handling for this type of implementation. Wouldn't that need to be coordinated with the networking layer? Yes. Doesn't it do the same thing, matching TX/RX queues to CPUs? Yes. If so, wouldn't we decrease bandwidth by restricting things to one CPU? We have a session or connection per CPU though, so we end up hitting the same problem you talked about last year where one hctx (iscsi session or connection's socket or nic hw queue) could get overloaded. This is what I meant in my original mail where iscsi would rely on whatever blk/mq load balancers we end up implementing at that layer to balance requests across hctxs. -- You received this message because you are subscribed to the Google Groups open-iscsi group. To unsubscribe from this group and stop receiving emails from it, send an email to open-iscsi+unsubscr...@googlegroups.com. To post to this group, send email to open-iscsi@googlegroups.com. Visit this group at http://groups.google.com/group/open-iscsi. For more options, visit https://groups.google.com/d/optout.
Re: [Lsf-pc] [LSF/MM TOPIC] iSCSI MQ adoption via MCS discussion
On Thu, 2015-01-08 at 14:57 -0800, Nicholas A. Bellinger wrote: On Thu, 2015-01-08 at 14:29 -0800, James Bottomley wrote: On Thu, 2015-01-08 at 14:16 -0800, Nicholas A. Bellinger wrote: On Thu, 2015-01-08 at 08:50 +0100, Bart Van Assche wrote: On 01/07/15 22:39, Mike Christie wrote: On 01/07/2015 10:57 AM, Hannes Reinecke wrote: On 01/07/2015 05:25 PM, Sagi Grimberg wrote: Hi everyone, Now that scsi-mq is fully included, we need an iSCSI initiator that would use it to achieve scalable performance. The need is even greater for iSCSI offload devices and transports that support multiple HW queues. As iSER maintainer I'd like to discuss the way we would choose to implement that in iSCSI. My measurements show that iSER initiator can scale up to ~2.1M IOPs with multiple sessions but only ~630K IOPs with a single session where the most significant bottleneck the (single) core processing completions. In the existing single connection per session model, given that command ordering must be preserved session-wide, we end up in a serial command execution over a single connection which is basically a single queue model. The best fit seems to be plugging iSCSI MCS as a multi-queued scsi LLDD. In this model, a hardware context will have a 1x1 mapping with an iSCSI connection (TCP socket or a HW queue). iSCSI MCS and it's role in the presence of dm-multipath layer was discussed several times in the past decade(s). The basic need for MCS is implementing a multi-queue data path, so perhaps we may want to avoid doing any type link aggregation or load balancing to not overlap dm-multipath. For example we can implement ERL=0 (which is basically the scsi-mq ERL) and/or restrict a session to a single portal. As I see it, the todo's are: 1. Getting MCS to work (kernel + user-space) with ERL=0 and a round-robin connection selection (per scsi command execution). 2. Plug into scsi-mq - exposing num_connections as nr_hw_queues and using blk-mq based queue (conn) selection. 3. Rework iSCSI core locking scheme to avoid session-wide locking as much as possible. 4. Use blk-mq pre-allocation and tagging facilities. I've recently started looking into this. I would like the community to agree (or debate) on this scheme and also talk about implementation with anyone who is also interested in this. Yes, that's a really good topic. I've pondered implementing MC/S for iscsi/TCP but then I've figured my network implementation knowledge doesn't spread that far. So yeah, a discussion here would be good. Mike? Any comments? I have been working under the assumption that people would be ok with MCS upstream if we are only using it to handle the issue where we want to do something like have a tcp/iscsi connection per CPU then map the connection to a blk_mq_hw_ctx. In this more limited MCS implementation there would be no iscsi layer code to do something like load balance across ports or transport paths like how dm-multipath does, so there would be no feature/code duplication. For balancing across hctxs, then the iscsi layer would also leave that up to whatever we end up with in upper layers, so again no feature/code duplication with upper layers. So pretty non controversial I hope :) If people want to add something like round robin connection selection in the iscsi layer, then I think we want to leave that for after the initial merge, so people can argue about that separately. Hello Sagi and Mike, I agree with Sagi that adding scsi-mq support in the iSER initiator would help iSER users because that would allow these users to configure a single iSER target and use the multiqueue feature instead of having to configure multiple iSER targets to spread the workload over multiple cpus at the target side. And I agree with Mike that implementing scsi-mq support in the iSER initiator as multiple independent connections probably is a better choice than MC/S. RFC 3720 namely requires that iSCSI numbering is session-wide. This means maintaining a single counter for all MC/S sessions. Such a counter would be a contention point. I'm afraid that because of that counter performance on a multi-socket initiator system with a scsi-mq implementation based on MC/S could be worse than with the approach with multiple iSER targets. Hence my preference for an approach based on multiple independent iSER connections instead of MC/S. The idea that a simple session wide counter for command sequence number assignment adds such a degree of contention that it
Re: [Lsf-pc] [LSF/MM TOPIC] iSCSI MQ adoption via MCS discussion
On Wed 07-01-15 09:22:13, Lee Duncan wrote: On 01/07/2015 08:25 AM, Sagi Grimberg wrote: Hi everyone, Now that scsi-mq is fully included, we need an iSCSI initiator that would use it to achieve scalable performance. The need is even greater for iSCSI offload devices and transports that support multiple HW queues. As iSER maintainer I'd like to discuss the way we would choose to implement that in iSCSI. My measurements show that iSER initiator can scale up to ~2.1M IOPs with multiple sessions but only ~630K IOPs with a single session where the most significant bottleneck the (single) core processing completions. In the existing single connection per session model, given that command ordering must be preserved session-wide, we end up in a serial command execution over a single connection which is basically a single queue model. The best fit seems to be plugging iSCSI MCS as a multi-queued scsi LLDD. In this model, a hardware context will have a 1x1 mapping with an iSCSI connection (TCP socket or a HW queue). iSCSI MCS and it's role in the presence of dm-multipath layer was discussed several times in the past decade(s). The basic need for MCS is implementing a multi-queue data path, so perhaps we may want to avoid doing any type link aggregation or load balancing to not overlap dm-multipath. For example we can implement ERL=0 (which is basically the scsi-mq ERL) and/or restrict a session to a single portal. As I see it, the todo's are: 1. Getting MCS to work (kernel + user-space) with ERL=0 and a round-robin connection selection (per scsi command execution). 2. Plug into scsi-mq - exposing num_connections as nr_hw_queues and using blk-mq based queue (conn) selection. 3. Rework iSCSI core locking scheme to avoid session-wide locking as much as possible. 4. Use blk-mq pre-allocation and tagging facilities. I've recently started looking into this. I would like the community to agree (or debate) on this scheme and also talk about implementation with anyone who is also interested in this. Cheers, Sagi. I started looking at this last year (and Hannes' suggestion), and would love to join the discussion. Please add me to the list of those that wish to attend. For that please send a separate email with attend request as described in the call for proposals. Thanks! Honza -- Jan Kara j...@suse.cz SUSE Labs, CR -- You received this message because you are subscribed to the Google Groups open-iscsi group. To unsubscribe from this group and stop receiving emails from it, send an email to open-iscsi+unsubscr...@googlegroups.com. To post to this group, send email to open-iscsi@googlegroups.com. Visit this group at http://groups.google.com/group/open-iscsi. For more options, visit https://groups.google.com/d/optout.
Re: [Lsf-pc] [LSF/MM TOPIC] iSCSI MQ adoption via MCS discussion
On Thu, 2015-01-08 at 21:03 -0800, Nicholas A. Bellinger wrote: On Thu, 2015-01-08 at 15:22 -0800, James Bottomley wrote: On Thu, 2015-01-08 at 14:57 -0800, Nicholas A. Bellinger wrote: On Thu, 2015-01-08 at 14:29 -0800, James Bottomley wrote: On Thu, 2015-01-08 at 14:16 -0800, Nicholas A. Bellinger wrote: SNIP The point is that a simple session wide counter for command sequence number assignment is significantly less overhead than all of the overhead associated with running a full multipath stack atop multiple sessions. I don't see how that's relevant to issue speed, which was the measure we were using: The layers above are just a hopper. As long as they're loaded, the MQ lower layer can issue at full speed. So as long as the multipath hopper is efficient enough to keep the queues loaded there's no speed degradation. The problem with a sequence point inside the MQ issue layer is that it can cause a stall that reduces the issue speed. so the counter sequence point causes a degraded issue speed over the multipath hopper approach above even if the multipath approach has a higher CPU overhead. Now, if the system is close to 100% cpu already, *then* the multipath overhead will try to take CPU power we don't have and cause a stall, but it's only in the flat out CPU case. Not to mention that our iSCSI/iSER initiator is already taking a session wide lock when sending outgoing PDUs, so adding a session wide counter isn't adding any additional synchronization overhead vs. what's already in place. I'll leave it up to the iSER people to decide whether they're redoing this as part of the MQ work. Session wide command sequence number synchronization isn't something to be removed as part of the MQ work. It's a iSCSI/iSER protocol requirement. The sequence number is a requirement of the session. Multiple separate sessions means no SN correlation between the different connections, so no global requirement for a SN counter across the queues ... that's what Mike was saying about implementing multipath not using MCS. With MCS we have a single session for all the queues and thus have to correlate the sequence number across all the connections and hence all the queues; without it we don't. That's why the sequence number becomes a potential stall point in MQ implementation of MCS which can be obviated if we use a separate session per queue. James -- You received this message because you are subscribed to the Google Groups open-iscsi group. To unsubscribe from this group and stop receiving emails from it, send an email to open-iscsi+unsubscr...@googlegroups.com. To post to this group, send email to open-iscsi@googlegroups.com. Visit this group at http://groups.google.com/group/open-iscsi. For more options, visit https://groups.google.com/d/optout.