Description of problem: While running Heketi on RHEL6, adding a device to a Heketi cluster crashes the Heketi server. Version-Release number of selected component (if applicable): heketi-1.0.2-1.el6rhs.x86_64 How reproducible: Always Steps to Reproduce: 1. Create a cluster, add a node, and then try add a device to the node 2. 3. Actual results: # heketi-cli device add -name=/dev/vdb -node=78cc8391d136fbaafc45e8c0c15563d6 Error: Get http://localhost:8081/queue/b7964d57ce86a1b68c2fb662ff54fb18: read tcp 127.0.0.1:8081: connection reset by peer Additional info: From heketi log: [heketi] INFO 2016/01/05 02:52:56 Adding device /dev/vdb to node 78cc8391d136fbaafc45e8c0c15563d6 [negroni] Completed 202 Accepted in 1.867598ms [asynchttp] INFO 2016/01/05 02:52:56 Started job b7964d57ce86a1b68c2fb662ff54fb18 panic: runtime error: invalid memory address or nil pointer dereference [signal 0xb code=0x1 addr=0x0 pc=0x5ba2e0] goroutine 8 [running]: github.com/heketi/heketi/executors/sshexec.(*SshExecutor).accessConnection(0x0, 0xc208096ea0, 0x1f) /builddir/build/BUILD/heketi-6563551111f7178b679866e85a0682325929d037/src/github.com/heketi/heketi/executors/sshexec/sshexec.go:97 +0x150 github.com/heketi/heketi/executors/sshexec.(*SshExecutor).sshExec(0x0, 0xc208096ea0, 0x1f, 0xc2080afd40, 0x2, 0x2, 0x5, 0x0, 0x0, 0x0, ...) /builddir/build/BUILD/heketi-6563551111f7178b679866e85a0682325929d037/src/github.com/heketi/heketi/executors/sshexec/sshexec.go:118 +0x89 github.com/heketi/heketi/executors/sshexec.(*SshExecutor).DeviceSetup(0x0, 0xc208096ea0, 0x1f, 0xc2080a4a70, 0x8, 0xc208097320, 0x20, 0x0, 0x0, 0x0) /builddir/build/BUILD/heketi-6563551111f7178b679866e85a0682325929d037/src/github.com/heketi/heketi/executors/sshexec/device.go:48 +0x470 github.com/heketi/heketi/apps/glusterfs.func·015(0x0, 0x0, 0x0, 0x0) /builddir/build/BUILD/heketi-6563551111f7178b679866e85a0682325929d037/src/github.com/heketi/heketi/apps/glusterfs/app_device.go:72 +0x159 github.com/heketi/heketi/rest.func·001() /builddir/build/BUILD/heketi-6563551111f7178b679866e85a0682325929d037/src/github.com/heketi/heketi/rest/asynchttp.go:128 +0x157 created by github.com/heketi/heketi/rest.(*AsyncHttpManager).AsyncHttpRedirectFunc /builddir/build/BUILD/heketi-6563551111f7178b679866e85a0682325929d037/src/github.com/heketi/heketi/rest/asynchttp.go:138 +0xfb goroutine 1 [select]: main.main() /builddir/build/BUILD/heketi-6563551111f7178b679866e85a0682325929d037/main.go:153 +0x1463
I'll take a look.
This is really weird. The panic points to https://github.com/heketi/heketi/blob/master/executors/sshexec/sshexec.go#L97 which cannot be null. It really is starting to feel like a compiler bug, but before I continue, I will try to reproduce. It looks like rhel6 is using golang 1.3.3.
I reinstalled my RHEL6 VM, and now I am unable to reproduce it either. Will keep this bug open for a while until I try it for a few more times.
I haven't been able to reproduce the issue on a fresh RHEL6 machine. Closing this bug.