On Tue, Oct 8, 2013 at 5:34 PM, Caspar Smit <[email protected]> wrote: > Hi Rasto, > > I have a relatively large list of issues with the latest LCMC >
Hi Caspar, some of the issues I was able to fix for the next 1.5.12 release > > 1) When I use the Cluster Wizard to create a new cluster and enter a name > for it (e.g. "Test") the default name of "Cluster 1" will always be used. > Only when I quit LCMC and restart it the cluster will be named "Test" fixed > > 2) When I start LCMC with an incorrect drbd config (for instance a wrong IP > address) i get the following error (which is expected): > > drbd config error: drbd.d/r0.res:8: in resource r0, on node01: > IP X.X.X.X not found on this host. > > But then I would like to edit this IP address in LCMC but the whole config > of r0 is not read and is not shown in LCMC so it is not editable either. > fixed > 3) Under DRBD the "DRBD Processes" button always results in error > > I noticed that LCMC tries to run a proxy command (drbd-proxy-ctl) during > this button press so I guess that's the reason for the error: > > sudo: clususer :command not allowed ; TTY=pts/7 ; PWD=/home/clususer ; > USER=root ; COMMAND=/sbin/drbd-proxy-ctl -c show hconnections -c show > hsubconnections -c show memusage -c print statistics > > I don't have DRBD proxy on this machine but I think this should not result > in an error. > I'll fix that > 4) When I create a VG from a newly created DRBD volume the VG is not visible > to LCMC even after a "Rescan LVM" > > I have to disconnect / reconnect to the cluster to see the newly created VG > and the "Create LV in VG" option. Rescan LVM doesn't seem to work at all. > Since you are using sudo, I suppose the lvdisplay, vgdisplay and pgdisplay commands are not in your sudoers config. Otherwise it should work. > 5) When creating a VM the "Browse" button for the installation disk is > greyed out when running LCMC from Windows (Ubuntu/Linux is working > correctly) Yup, in this case the implementation is platform dependent and I haven't got around to do it for Windows. > > 6) It would be nice to have "Driver Cache" option while creating a VM in the > "Storage" section, the Driver Cache setting is only available in the > Installation Disk section. fixed > > 7) If during the VM creation wizard in the Storage section a Virtio disk is > added, LCMC still adds this virtio disk as "hda" in stead of "vda" fixed > > 8) When loading the DRBD module through LCMC and I connect a resource i > often get an error message on the console saying: > > Oct 8 16:49:16 node01 kernel: [ 6926.447449] block drbd0: Different > verify-alg settings. me="crc32c" peer="" > > Although the config was created through LCMC and is equal on both sides. > > A "Load Config (Adjust)" resolves the issue but it shouldn't happen in the > first place. > I am not sure at which place does this happen. If you just load the module without adjust, you shouldn't be able to connect the resources. Or do you mean that you load the module and then create a new DRBD resource? Can you write up the steps to reproduce it? Also, what DRBD version? > 9) I mailed to this list some time ago about the /dev/disk/by-id but i > didn't got any response so here is it again: It's on the TODO list, but i am not planning it for the next release. > > > 10) When adding a CRM resource (linbit:drbd) and you give the master/slave > set a different name then the default (Which is drbd_1) it still gets added > to the CIB with the name ms_drbd_1 even if you give it a different name like > ms_drbd0 for instance. > > You have to disconnect/reconnect for LCMC to see the difference and be able > to manage the newly created resource. > > ms ms_drbd_1 res_drbd_drbd0 \ > meta clone-max="2" notify="true" interleave="true" fixed > > 11) When there is no *.res file in /etc/drbd.d or there is no drbd resource > configured the global_common.conf file is not parsed. So when you create a > drbd volume using LCMC the global_common.conf file is overwritten with very > little options set. I would like to keep my global_common.conf file > allthough there is no drbd volume configured (yet). > > I don't want to manually set all these options again and LCMC should just > overwrite it. I've played with it a little bit. Eventually the drbdadm would have to be patched. Here's the patch, I think :) >From a38ed73550a2044805726e208e34a4f628301eed Mon Sep 17 00:00:00 2001 From: Rasto Levrinc <[email protected]> Date: Mon, 14 Oct 2013 23:32:04 +0200 Subject: [PATCH] drbdadm: allow dump/-xml without configured resources --- user/drbdadm_main.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/user/drbdadm_main.c b/user/drbdadm_main.c index 170e71f..5f60d4e 100644 --- a/user/drbdadm_main.c +++ b/user/drbdadm_main.c @@ -1012,6 +1012,9 @@ static int adm_dump(struct cfg_ctx *ctx) struct d_host_info *host; struct d_resource *res = ctx->res; + if (res == NULL) + return 0; + printI("# resource %s on %s: %s, %s\n", esc(res->name), nodeinfo.nodename, res->ignore ? "ignored" : "not ignored", @@ -1042,6 +1045,9 @@ static int adm_dump_xml(struct cfg_ctx *ctx) struct d_host_info *host; struct d_resource *res = ctx->res; + if (res == NULL) + return 0; + printI("<resource name=\"%s\" conf-file-line=\"%s:%u\">\n", esc_xml(res->name), esc_xml(res->config_file), res->start_line); @@ -4044,7 +4050,7 @@ int main(int argc, char **argv) ctx.arg = cmd->name; if (cmd->res_name_required) { - if (config == NULL) { + if (config == NULL && !is_dump) { fprintf(stderr, "no resources defined!\n"); exit(E_usage); } > > As i said, a fairly big list. I hope you can confirm and reproduce these > issues so they can be fixed Thanks, the bigger the better, Rasto > > Thanks in advance and keep up the good work on LCMC! > > Kind regards, > Caspar Smit _______________________________________________ drbd-mc mailing list [email protected] http://lists.linbit.com/mailman/listinfo/drbd-mc
