[dm-devel] [PATCH V4 1/2] multipath-tools: intermittent IO error accounting to improve reliability

[Muneendra Kumar M]

Hi Guan,
>>>Shall we use existing PATH_SHAKY ?
As the path_shaky Indicates path not available for "normal" operations we can use this state. That's  a good idea.

Regarding the marginal paths below is my explanation. And brocade is publishing couple of white papers regarding the same to educate the SAN administrators and the san community.

Marginal path:

A host, target, LUN (ITL path) flow  goes through SAN. It is to be noted that the for each I/O request that goes to the SCSI layer, it transforms into a single SCSI exchange.  In a single SAN, there are typically multiple SAN network paths  for a ITL flow/path. Each SCSI exchange  can take one of the various network paths that are available for the ITL path.  A SAN can be based on Ethernet, FC, Infiniband physical networks to carry block storage traffic (SCSI, NVMe etc.)

There are typically two type of SAN network problems that are categorized as marginal issues. These issues by nature are not permanent in time and do come and go away over time.
1) Switches in the SAN can have intermittent frame drops or intermittent frame corruptions due to bad optics cable (SFP) or any such wear/tear  port issues. This causes ITL flows that go through the faulty switch/port to intermittently experience frame drops.  
2) There exists SAN topologies where there are switch ports in the fabric that becomes the only  conduit for many different ITL flows across multiple hosts. These single network paths are essentially shared across multiple ITL flows. Under these conditions if the port link bandwidth is not able to handle the net sum of the shared ITL flows bandwidth going through the single path  then we could see intermittent network congestion problems. This condition is called network oversubscription. The intermittent congestions can delay SCSI exchange completion time (increase in I/O latency is observed).

To overcome the above network issues and many more such target issues, there are frame level retries that are done in HBA device firmware and I/O retries in the SCSI layer. These retries might succeed because of two reasons:
1) The intermittent switch/port issue is not observed
2) The retry I/O is a new  SCSI exchange. This SCSI exchange  can take an alternate SAN path for the ITL flow, if such an SAN path exists.
3) Network congestion disappears momentarily because the net I/O bandwidth coming from multiple ITL flows on the single shared network path is something the path can handle

However in some cases we have seen I/O retries don’t succeed because the retry I/Os hits a SAN network path that has  intermittent switch/port issue and/or network congestion. 

On the host  thus we see configurations two or more ITL path sharing the same target/LUN going through two or more HBA ports. These HBA ports are connected to two or more SAN to the same target/LUN.
If the I/O fails at the multipath layer then, the ITL path is turned into Failed state. Because of the marginal nature of the network, the next Health Check command sent from multipath layer might succeed, which results in making the ITL path into Active state. You end up seeing the DM path state going  into Active, Failed, Active transitions. This results in overall reduction in application I/O throughput and sometime application I/O failures (because of timing constraints). All this can happen because of I/O retries and I/O request moving across multiple paths of the DM device. In the host it is  to be noted all I/O retries on a single path and I/O movement across multiple paths results in slowing down the forward progress of new application I/O. Reason behind, the above I/O  re-queue actions are given higher priority than the newer I/O requests coming from the application. 

The above condition of the  ITL path is hence called “marginal”.

What we desire is for the DM to deterministically  categorize a ITL Path as “marginal” and move all the pending I/Os from the marginal Path to an Active Path. This will help in meeting application I/O timing constraints. Also a capability to automatically re-instantiate the marginal path into Active once the marginal condition in the network is fixed.


Based on the above explanation I want to rename the names as marginal_path_XXXX and this is irrespective of any storage network.

Regards,
Muneendra.



-----Original Message-----
From: Guan Junxiong [mailto:guanjunxiong at huawei.com] 
Sent: Tuesday, September 19, 2017 6:23 PM
To: Muneendra Kumar M <mmandala at Brocade.com>; Martin Wilck <mwilck at suse.com>; dm-devel at redhat.com; christophe.varoqui at opensvc.com
Cc: shenhong09 at huawei.com; niuhaoxin at huawei.com; chengjike.cheng at huawei.com
Subject: Re: [PATCH V4 1/2] multipath-tools: intermittent IO error accounting to improve reliability

Hi Muneendra ,
Thanks for your suggestion. My comments inline.

On 2017/9/19 18:59, Muneendra Kumar M wrote:
> Hi Guan/Martin,
> Below are my points.
> 
>>>> "san_path_double_fault_time"  is great.  One less additional parameter and still covering most scenarios are appreciated.
> This looks good and I completely agree with Guan.
> 
> One question san_path_double_fault_time is the time between two failed states (failed-active-failed )is this correct ?.
> Then this holds good.
> 
> Instead of san_path_double_fault_time can we call it as san_path_double_failed_time as the name suggests the time between two failed states is this ok ?
> 

Both names are fine for me.

> In SAN  topology (FC,NVME,SCSI)transient intermittent network errors make the ITL paths  as marginal paths. 
>
>
> So instead of calling "path_io_err_sample_time", "path_io_err_rate_threshold" and "path_io_err_recovery_time"
> can we name as "marginal_path_err_detection_time", " marginal_path_err_rate_threshold" and " marginal_path_err_recovery_time"
> 
> Some other names should also be good as the io_path is general word from my view.
> >

Can you explain "the marginal paths" in details?  Can the users easily catch the meaning of  marginal paths?
IMO, path_io_err_XXXs are easy to under_stand

> If we agree with this one more thing which I would like to add as part of this patch.
> 
> Whenever the path is in XXX_io_error_recovery_time  and if the user runs multipath -ll command the path the state of the path is shown as failed as shown below.
> 
> 	| `- 6:0:0:0 sdb 8:16  failed ready  running
> 
> Can we add a new state as marginal so that when the admin run the 
> multipath command and found that the state is in marginal he can quickly come to know that this a marginal Path and needs to be recovered .If we keep the state as failed the admin cannot understand from past how much time the device is in failed state.
> 
>

If the user use multipathd -k , then input "show paths", the multipathd will show the path in the state of "delayed".
But we can't find the exactly reason of the delayed state path because other features such as path waiting uses it.
Shall we use existing PATH_SHAKY ?

Regards.
Guan