<div dir="ltr">I have isolated the issue to be related to readdirp. Both the clients where I see the messages are running relatively newer kernels (3.13+), and by default the kernel readdirp support is enabled. If I mount with use-readdirp=no, I don't see the Warning messages. Is this a known issue ?</div><div class="gmail_extra"><br><div class="gmail_quote">On Mon, Sep 29, 2014 at 3:28 AM, Prasun Gera <span dir="ltr"><<a href="mailto:prasun.gera@gmail.com" target="_blank">prasun.gera@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr"><div>Hello,</div>I set up a distributed storage on 3 nodes using glusterfs, and on a couple of my clients, I am seeing a lot of these messages<div><br></div><div><br></div><div><div><div>[2014-09-29 06:34:16.549972] W [client-rpc-fops.c:2774:client3_3_lookup_cbk] 0-hparch_gluster-client-2: remote operation failed: No such file or directory. Path: /trace/1/spec2006/trace_simpoint/bzip2/bbdata/bzip21.T.0.bb.simpoints.lpt0.9999 (874702dc-058a-4871-8c9f-11289c396af5)</div><div>[2014-09-29 06:34:16.550934] W [client-rpc-fops.c:2774:client3_3_lookup_cbk] 0-hparch_gluster-client-1: remote operation failed: No such file or directory. Path: /trace/1/spec2006/trace_simpoint/bzip2/bbdata/bzip21.T.0.bb.simpoints.lpt0.9999 (874702dc-058a-4871-8c9f-11289c396af5)</div></div></div><div>.....</div><div><br></div><div><br></div><div>I am able to access these files though. Before these warning messages start, there are informational messages which mention difference between the server and the client versions. </div><div><br></div><div><div>[2014-09-29 05:36:25.117971] I [client-handshake.c:1677:select_server_supported_programs] 0-hparch_gluster-client-2: Using Program GlusterFS 3.3, Num (1298437), Version (330)</div><div>[2014-09-29 05:36:25.122228] I [client-handshake.c:1462:client_setvolume_cbk] 0-hparch_gluster-client-2: Connected to <a href="http://143.215.128.92:49153" target="_blank">143.215.128.92:49153</a>, attached to remote volume '/bricks/hparch'.</div><div><div>[2014-09-29 05:36:27.932384] I [client-handshake.c:1474:client_setvolume_cbk] 0-hparch_gluster-client-0: Server and Client lk-version numbers are not same, reopening the fds</div></div><div><br></div><div>.......</div></div><div><br></div><div><br></div><div>The servers are running RHSS2.1 which I guess is based on 3.4 and one of my clients is ubuntu and the other one is RHEL6. The ubuntu client is running the 3.5 ppa package and the rhel client is running the client included, which is at 3.6 now. The RHEL client also prints out messages about missing trash directories (in addition to these), which I guess is a 3.6 feature. Does anyone know what's going on ? Googling around pointed me to self-heal issues. I haven't run self heal and I don't even have a replicated setup. </div><div><br></div><div><br></div><div><br></div><div>Regards,</div><div>Prasun</div><div><br></div><div><br></div><div><br></div></div>
</blockquote></div><br></div>