Sent from YoMail <bjq1016> 2016-10-01 09:30:16 wrote: Hello everyone, First, I configured a smb cluster of 3 nodes with ctdb . I found that when smb is killed or stop nomarlly , it cannot be restarted by ctdb, but the public ip reallo
On 1 October 2016 12:14:10 pm bjq1016 via samba <samba at lists.samba.org> wrote:> <bjq1016> 2016-10-01 09:30:16 wrote: > Hello everyone, > First, I configured a smb cluster of 3 nodes with ctdb . I found > that when smb is killed or stop nomarlly , it cannot be restarted by ctdb, > but the public ip realloI think you're saying that CTDB doesn't try to restart smbd, CTDB notices smbd is down and reallocates public IP addresses. That's right. CTDB does try to restart other services. However, it doesn't restart Samba because we don't want Samba to flip-flop... peace & happiness, martin
Thank you. But what if all the smb on 3 nodes are down ? < Martin Schwenke> 2016-10-01 10:26:10 wrote: On 1 October 2016 12:14:10 pm bjq1016 via samba <samba at lists.samba.org> wrote:> <bjq1016> 2016-10-01 09:30:16 wrote: > Hello everyone, >
On 2016-10-01 at 13:34 +0800, bjq1016 via samba wrote:> > Thank you. But what if all the smb on 3 nodes are down ?If it's a 3-node cluster, and 3 nodes are down, then the ctdb/samba cluster is down (not reacheable via smb). The admin needs to investigate in that case. That's the design. No automatic restarting of smbd. To prevent endless loops. Michael -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 163 bytes Desc: not available URL: <http://lists.samba.org/pipermail/samba/attachments/20161001/67725f8e/signature.sig>