On Fri, Dec 2, 2016 at 8:23 AM, Xusangdi <xu.sangdi@xxxxxxx> wrote: > Hi John, > > In our environment we want to deploy MDS and cephfs client on the same node (users actually use cifs/nfs to access ceph storage). However, > it takes a long time to recover if the node with active MDS fails, during which a large part is for the new MDS waiting for all clients reconnect. > The `mds_reconnect_timeout` is set to 45s by default, but in our experiment cluster, the reconnection with normal client is usually quite fast (<100ms), > so imho we might configure it to a much smaller value like 5, but I’m not sure if that would be safe. Any suggestions? Is there any concern I missed > that inspires to such a large default value? If a client doesn't reconnect within the timeout, then it will be evicted (and any buffered writes it had might be lost). Evicting clients is kind of a big deal, so we only want to do it if we're really sure they're dead, hence this being quite a long timeout. However, if you know that a client is really gone and you don't want to wait for it, then you can use the "session evict" admin socket command to evict the client immediately when the MDS goes into the reconnect state, and avoid waiting. John > Regards, > --- Sandy > ------------------------------------------------------------------------------------------------------------------------------------- > 本邮件及其附件含有杭州华三通信技术有限公司的保密信息，仅限于发送给上面地址中列出 > 的个人或群组。禁止任何其他人以任何形式使用（包括但不限于全部或部分地泄露、复制、 > 或散发）本邮件中的信息。如果您错收了本邮件，请您立即电话或邮件通知发件人并删除本 > 邮件！ > This e-mail and its attachments contain confidential information from H3C, which is > intended only for the person or entity whose address is listed above. Any use of the > information contained herein in any way (including, but not limited to, total or partial > disclosure, reproduction, or dissemination) by persons other than the intended > recipient(s) is prohibited. If you receive this e-mail in error, please notify the sender > by phone or email immediately and delete it! _______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
 |