Re: [PATCH RFC] svcrdma: Ignore source port when computing DRC hash

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 6/10/2019 6:13 PM, Tom Talpey wrote:
On 6/10/2019 5:57 PM, Chuck Lever wrote:


On Jun 10, 2019, at 3:14 PM, Tom Talpey <tom@xxxxxxxxxx> wrote:

On 6/10/2019 1:50 PM, Chuck Lever wrote:
Hi Tom-
On Jun 10, 2019, at 10:50 AM, Tom Talpey <tom@xxxxxxxxxx> wrote:

On 6/5/2019 1:25 PM, Chuck Lever wrote:
Hi Tom-
On Jun 5, 2019, at 12:43 PM, Tom Talpey <tom@xxxxxxxxxx> wrote:

On 6/5/2019 8:15 AM, Chuck Lever wrote:
The DRC is not working at all after an RPC/RDMA transport reconnect. The problem is that the new connection uses a different source port,
which defeats DRC hash.

An NFS/RDMA client's source port is meaningless for RDMA transports.
The transport layer typically sets the source port value on the
connection to a random ephemeral port. The server already ignores it
for the "secure port" check. See commit 16e4d93f6de7 ("NFSD: Ignore
client's source port on RDMA transports").

Where does the entropy come from, then, for the server to not
match other requests from other mount points on this same client?
The first ~200 bytes of each RPC Call message.
[ Note that this has some fun ramifications for calls with small
RPC headers that use Read chunks. ]

Ok, good to know. I forgot that the Linux server implemented this.
I have some concerns abot it, honestly, and it's important to remember
that it's not the same on all servers. But for the problem you're
fixing, it's ok I guess and certainly better than today. Still, the
errors are goingto be completely silent, and can lead to data being
corrupted. Well, welcome to the world of NFSv3.
I don't see another option.
Some regard this checksum as more robust than using the client's
IP source port. After all, the same argument can be made that
the server cannot depend on clients to reuse their source port.
That is simply a convention that many clients adopted before
servers used a stronger DRC hash mechanism.
And since RDMA is capable of
such high IOPS, the likelihood seems rather high.
Only when the server's durable storage is slow enough to cause
some RPC requests to have extremely high latency.
And, most clients use an atomic counter for their XIDs, so they
are also likely to wrap that counter over some long-pending RPC
request.
The only real answer here is NFSv4 sessions.
Missing the cache
might actually be safer than hitting, in this case.
Remember that _any_ retransmit on RPC/RDMA requires a fresh
connection, that includes NFSv3, to reset credit accounting
due to the lost half of the RPC Call/Reply pair.
I can very quickly reproduce bad (non-deterministic) behavior
by running a software build on an NFSv3 on RDMA mount point
with disconnect injection. If the DRC issue is addressed, the
software build runs to completion.

Ok, good. But I have a better test.

In the Connectathon suite, there's a "Special" test called "nfsidem".
I wrote this test in, like, 1989 so I remember it :-)

This test performs all the non-idempotent NFv3 operations in a loop,
and each loop element depends on the previous one, so if there's
any failure, the test imemdiately bombs.

Nobody seems to understand it, usually when it gets run people will
run it without injecting errors, and it "passes" so they decide
everything is ok.

So my suggestion is to run your flakeway packet-drop harness while
running nfsidem in a huge loop (nfsidem 10000). The test is slow,
owing to the expensive operations it performs, so you'll need to
run it for a long time.

You'll almost definitely get a failure or two, since the NFSv3
protocol is flawed by design. But you can compare the behaviors,
and even compute a likelihood. I'd love to see some actual numbers.
I configured the client to disconnect after 23711 RPCs have completed.
(I can re-run these with more frequent disconnects if you think that
would be useful).
Here's a run with the DRC modification:
[cel@manet ~]$ sudo mount -o vers=3,proto=rdma,sec=sys klimt.ib:/export/tmp /mnt
[cel@manet ~]$ (cd /mnt; ~/src/cthon04/special/nfsidem 100000)
testing 100000 idempotencies in directory "./TEST"
[cel@manet ~]$ sudo umount /mnt
Here's a run with the stock v5.1 Linux server:
[cel@manet ~]$ sudo mount -o vers=3,proto=rdma,sec=sys klimt.ib:/export/tmp /mnt
[cel@manet ~]$ (cd /mnt; ~/src/cthon04/special/nfsidem 100000)
testing 100000 idempotencies in directory "./TEST"
[cel@manet ~]$
This test reported no errors in either case. We can see that the
disconnects did trigger retransmits:
RPC statistics:
   1888819 RPC requests sent, 1888581 RPC replies received (0 XIDs not found)
   average backlog queue length: 119

Ok, well, that's 1.2% error rate, which IMO could be cranked up much
higher for testing purposes. I'd also be sure the server was serving
other workloads during the same time, putting at least some pressure
on the DRC. The op rate of a single nfsidem test is pretty low so I
doubt it's ever evicting anything.

Ideally, it would be best to
1) increase the error probability
2) run several concurrent nfsidem tests, on different connections
3) apply some other load to the server, e.g. several cthon basics

The idea being to actually get the needle off of zero and measure some
kind of difference. Otherwise it really isn't giving any information
apart from a slight didn't-break-it confidence. Honestly, I'm surprised
you couldn't evince a failure from stock. On paper, these results don't
actually tell us the patch is doing anything.

I boosted the disconnect injection rate to once every 11353 RPCs,
and mounted a second share with "nosharecache", running nfsidem on
both mounts. Both mounts are subject to disconnect injection.

With the current v5.2-rc Linux server, both nfsidem jobs fail within
30 seconds.

With my DRC fix applied, both jobs run to completion with no errors.
It takes more than an hour.

Great, that definitely proves it. It's relatively low risk, and
should go in.

I think it's worth thinking through the best way to address this
across multiple transports, especially RDMA, where the port and
address behaviors may differ from TCP/UDP. Perhaps tossing the whole
idea of using the 5-tuple should simply be set aside, after all
these years. Even though NFSv4.1 already has!

Thanks for humoring me with the extra testing, by the way :-)

Tom.

Here are the op metrics for one of the mounts during a run that
completes successfully:

RPC statistics:
   4091380 RPC requests sent, 4088143 RPC replies received (0 XIDs not found)
   average backlog queue length: 1800

ACCESS:
            300395 ops (7%)         301 retrans (0%)        0 major timeouts
         avg bytes sent per op: 132    avg bytes received per op: 120
         backlog wait: 4.289199  RTT: 0.019821   total execute time: 4.315092 (milliseconds)
REMOVE:
            300390 ops (7%)         168 retrans (0%)        0 major timeouts
         avg bytes sent per op: 136    avg bytes received per op: 144
         backlog wait: 2.368664  RTT: 0.070106   total execute time: 2.445148 (milliseconds)
MKDIR:
           200262 ops (4%)         193 retrans (0%)        0 major timeouts
         avg bytes sent per op: 158    avg bytes received per op: 271
         backlog wait: 4.207034  RTT: 0.075421   total execute time: 4.289101 (milliseconds)
RMDIR:
           200262 ops (4%)         100 retrans (0%)        0 major timeouts
         avg bytes sent per op: 130    avg bytes received per op: 144
         backlog wait: 2.050749  RTT: 0.071676   total execute time: 2.128801 (milliseconds)
LOOKUP:
            194664 ops (4%)         233 retrans (0%)        0 major timeouts
         avg bytes sent per op: 136    avg bytes received per op: 176
         backlog wait: 5.365984  RTT: 0.020615   total execute time: 5.392769 (milliseconds)
SETATTR:
            100130 ops (2%)         86 retrans (0%)         0 major timeouts
         avg bytes sent per op: 160    avg bytes received per op: 144
         backlog wait: 3.520603  RTT: 0.066863   total execute time: 3.594327 (milliseconds)
WRITE:
            100130 ops (2%)         82 retrans (0%)         0 major timeouts
         avg bytes sent per op: 180    avg bytes received per op: 136
         backlog wait: 3.331249  RTT: 0.118316   total execute time: 3.459563 (milliseconds)
CREATE:
            100130 ops (2%)         95 retrans (0%)         0 major timeouts
         avg bytes sent per op: 168    avg bytes received per op: 272
         backlog wait: 4.099451  RTT: 0.071437   total execute time: 4.177479 (milliseconds)
SYMLINK:
            100130 ops (2%)         83 retrans (0%)         0 major timeouts
         avg bytes sent per op: 188    avg bytes received per op: 271
         backlog wait: 3.727704  RTT: 0.073534   total execute time: 3.807700 (milliseconds)
RENAME:
            100130 ops (2%)         68 retrans (0%)         0 major timeouts
         avg bytes sent per op: 180    avg bytes received per op: 260
         backlog wait: 2.659982  RTT: 0.070518   total execute time: 2.738979 (milliseconds)
LINK:
    100130 ops (2%)     85 retrans (0%)     0 major timeouts
    avg bytes sent per op: 172    avg bytes received per op: 232
    backlog wait: 3.676680     RTT: 0.066773     total execute time: 3.749785 (milliseconds)
GETATTR:
    230 ops (0%)     81 retrans (35%)     0 major timeouts
    avg bytes sent per op: 170    avg bytes received per op: 112
    backlog wait: 1584.026087     RTT: 0.043478     total execute time: 1584.082609 (milliseconds)

By the way, that's a *nasty* tail latency on at least one getattr.
What's up with that??

Tom.

READDIRPLUS:
    10 ops (0%)
    avg bytes sent per op: 149    avg bytes received per op: 1726
    backlog wait: 0.000000     RTT: 0.300000     total execute time: 0.400000 (milliseconds)
FSINFO:
    2 ops (0%)
    avg bytes sent per op: 112    avg bytes received per op: 80
    backlog wait: 0.000000     RTT: 0.000000     total execute time: 0.000000 (milliseconds)
NULL:
    1 ops (0%)
    avg bytes sent per op: 40    avg bytes received per op: 24
    backlog wait: 2.000000     RTT: 0.000000     total execute time: 3.000000 (milliseconds)
READLINK:
    1 ops (0%)
    avg bytes sent per op: 128    avg bytes received per op: 1140
    backlog wait: 0.000000     RTT: 0.000000     total execute time: 0.000000 (milliseconds)
PATHCONF:
    1 ops (0%)
    avg bytes sent per op: 112    avg bytes received per op: 56
    backlog wait: 0.000000     RTT: 0.000000     total execute time: 0.000000 (milliseconds)


Tom.

ACCESS:
         300001 ops (15%)        44 retrans (0%)         0 major timeouts
         avg bytes sent per op: 132    avg bytes received per op: 120
         backlog wait: 0.591118  RTT: 0.017463   total execute time: 0.614795 (milliseconds)
REMOVE:
            300000 ops (15%)        40 retrans (0%)         0 major timeouts
         avg bytes sent per op: 136    avg bytes received per op: 144
         backlog wait: 0.531667  RTT: 0.018973   total execute time: 0.556927 (milliseconds)
MKDIR:
          200000 ops (10%)        26 retrans (0%)         0 major timeouts
         avg bytes sent per op: 158      avg bytes received per op: 272
         backlog wait: 0.518940  RTT: 0.019755   total execute time: 0.545230 (milliseconds)
RMDIR:
    200000 ops (10%)        24 retrans (0%)         0 major timeouts
         avg bytes sent per op: 130    avg bytes received per op: 144
         backlog wait: 0.512320  RTT: 0.018580   total execute time: 0.537095 (milliseconds)
LOOKUP:
            188533 ops (9%)         21 retrans (0%)         0 major timeouts
         avg bytes sent per op: 136    avg bytes received per op: 174
         backlog wait: 0.455925  RTT: 0.017721   total execute time: 0.480011 (milliseconds)
SETATTR:
         100000 ops (5%)         11 retrans (0%)         0 major timeouts
         avg bytes sent per op: 160    avg bytes received per op: 144
         backlog wait: 0.371960  RTT: 0.019470   total execute time: 0.398330 (milliseconds)
WRITE:
           100000 ops (5%)         9 retrans (0%)  0 major timeouts
         avg bytes sent per op: 180    avg bytes received per op: 136
         backlog wait: 0.399190  RTT: 0.022860   total execute time: 0.436610 (milliseconds)
CREATE:
            100000 ops (5%)         9 retrans (0%)  0 major timeouts
         avg bytes sent per op: 168    avg bytes received per op: 272
         backlog wait: 0.365290  RTT: 0.019560   total execute time: 0.391140 (milliseconds)
SYMLINK:
          100000 ops (5%)         18 retrans (0%)         0 major timeouts
         avg bytes sent per op: 188    avg bytes received per op: 272
         backlog wait: 0.750470  RTT: 0.020150   total execute time: 0.786410 (milliseconds)
RENAME:
          100000 ops (5%)         14 retrans (0%)         0 major timeouts
         avg bytes sent per op: 180    avg bytes received per op: 260
         backlog wait: 0.461650  RTT: 0.020710   total execute time: 0.489670 (milliseconds)
--
Chuck Lever

--
Chuck Lever










[Index of Archives]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Photo]     [Yosemite News]     [Yosemite Photos]     [Linux Kernel]     [Linux SCSI]     [XFree86]

  Powered by Linux