Hi Seena,

The errno -5 is EIO, but as you saw it converts to 500--which is
retryable (clients should re-try).  As Yehuda said, the client could
have stalled, but this error will also be generated if the client
disconnects (e.g., timed out because RGW took too long).  The logs
should give a better indication what actually happened.

Matt

On Mon, Mar 23, 2020 at 9:00 AM Seena Fallah <seenafallah@gmail.com> wrote:
>
> If it’s because of client why the status code changed to 500 (server error)? Isn’t it a client error?
>
> On Mon, Mar 23, 2020 at 5:26 PM Yehuda Sadeh-Weinraub <yehuda@redhat.com> wrote:
>>
>> On Mon, Mar 23, 2020 at 2:28 PM Seena Fallah <seenafallah@gmail.com> wrote:
>> >
>> > Hi all.
>> > I often get this error on my PUT requests.
>> >
>> > rgw.log-552376-2020-03-23 15:12:53.270 7fe5aa6da700  0 WARNING: set_req_state_err err_no=5 resorting to 500
>> > rgw.log-552377-2020-03-23 15:12:53.270 7fe5aa6da700  1 ====== req done req=0x7fe5aa6d38c0 op status=-5 http_status=500 latency=59.9965s ======
>> >
>> > My put requests are always under 1s but this requests spend 1min!
>> > Any logs I can read to find out what's going on with my cluster?
>>
>> Likely that rgw waits for more data from the client and eventually
>> times out. This can happen due to various reasons. You can bump up rgw
>> logs (debug rgw = 20), and try to look at what the client thinks it's
>> sending vs what rgw receives. You can also try to look at tcp capture
>> (via wireshark for example).
>>
>> Yehuda
>>
> _______________________________________________
> Dev mailing list -- dev@ceph.io
> To unsubscribe send an email to dev-leave@ceph.io



--

Matt Benjamin
Red Hat, Inc.
315 West Huron Street, Suite 140A
Ann Arbor, Michigan 48103

http://www.redhat.com/en/technologies/storage

tel.  734-821-5101
fax.  734-769-8938
cel.  734-216-5309
_______________________________________________
Dev mailing list -- dev@ceph.io
To unsubscribe send an email to dev-leave@ceph.io