On Wed, Jun 27, 2018 at 12:47 PM, Mike Jumper <[email protected]> wrote: > On Wed, Jun 27, 2018 at 7:20 AM, Nicolas Lécureuil > <[email protected]> wrote: >> >> ... >> Jun 27 16:10:34 pulsemaster guacd[28941]: User is not responding. >> Jun 27 16:10:34 pulsemaster guacd[24552]: Connection >> "$12c5328c-c127-453d-bc9c-f631bf8e90f8" removed. >> Jun 27 16:10:34 pulsemaster kernel: [717204.421339] guacd[28944]: segfault >> at 10 ip 00007f9ead26c692 sp 00007f9eaec7ad10 error 4 in >> libguac-client-vnc.so.0.0.0[7f9ead264000+13000] >> ... >> > > This segfault is occurring after the connection has already ended, > while the resources consumed by that connection are being freed (the > "User is not responding" message indicates that the connection has > been closed due to a network timeout). I believe this was recently > fixed: > > https://issues.apache.org/jira/browse/GUACAMOLE-424 > > The segfault in the case of the above issue was due to an erroneous > attempt to free SFTP-related resources when SFTP was not actually > being used for the connection. If you're seeing unexpected > disconnections, the issue causing those disconnections is likely > elsewhere, as this particular bug did not affect connection stability. >
Unless I'm thinking of a different issue, and the above was related to freeing display resources that were not actually yet allocated. Either way ... if you're seeing unexpected disconnects, this segfault is due to something going wrong while handling the disconnect that has already occurred; it isn't the cause of the disconnect. - Mike
