|  | 
lvs-users
 | 
| To: | lvs-users@xxxxxxxxxxxxxxxxxxxxxx <lvs-users@xxxxxxxxxxxxxxxxxxxxxx> | 
|---|---|
| Subject: | Re: Re: anyone used director failover in production? | 
| From: | xiongqiuhua <mountains_0513@xxxxxxxx> | 
| Date: | Fri, 24 May 2002 9:41:50 +0800 | 
| hi,I am making preparations for a lvs handling mail server . can anyone who have the experience introduce something about it to me? 
 I have a production LVS server running with 3 real servers handling SMTP, POP3, IMAP for our QMAIL server.  We process about a million inbound connections a day.  I've never had the primary LVS server crash but I have shut it down on purpose (yanked the power cord) to test the fail over.  Everything worked perfectly. [root@lvsd-1 /root]# uname -a Linux lvsd-1.crocker.com 2.4.9 #1 Tue Aug 28 11:42:10 EDT 2001 i686 unknown [root@lvsd-1 /root]# uptime   5:27am  up 71 days, 16:47,  1 user,  load average: 1.06, 1.03, 1.00 [root@lvsd-1 /root]# /sbin/ipvsadm -L IP Virtual Server version 0.9.3 (size=65536) Prot LocalAddress:Port Scheduler Flags   -> 
 RemoteAddress:Port             Forward Weight ActiveConn InActConn TCP  mail.crocker.com:smtp lc   -> 
 cw-3-e0.crocker.com:smtp       Route   1      79         194   -> 
 cw-1-e0.crocker.com:smtp       Route   1      78         228   -> 
 cw-2-e0.crocker.com:smtp       Route   1      81         116 TCP  mail.crocker.com:pop3 lc   -> 
 cw-1-e0.crocker.com:pop3       Route   1      2          42   -> 
 cw-3-e0.crocker.com:pop3       Route   1      1          45   -> 
 cw-2-e0.crocker.com:pop3       Route   1      1          44 TCP  mail.crocker.com:imap2 lc   -> 
 cw-1-e0.crocker.com:imap2      Route   1      3          2   -> 
 cw-3-e0.crocker.com:imap2      Route   1      2          3   -> 
 cw-2-e0.crocker.com:imap2      Route   1      3          1 On Thu, 2002-05-23 at 06:58, Joseph Mack wrote: >> 
 From what I can tell (small number of anecdotes), director >> 
 failover is mainly used for planned maintenance and so that >> 
 management can tell customers that their setup has redundancy. >> 
 I haven't heard any stories of failover working in production. >>   >> 
 Does anyone have director failover in a production setup, when >> 
 directors have failed and changed over automatically and you've >> 
 only found out about it when the failover mechanism notified >> 
 you (rather than the customers notifying you)? >>   >> 
 If so, I'd like to put the info into the HOWTO. Please let >> 
 me know what failed, what system you're using for failover, >>  and anything else of interest. >>   >>  Thanks Joe >>   >>  --  >> 
 Joseph Mack PhD, Senior Systems Engineer, Lockheed Martin >> 
 contractor to the National Environmental Supercomputer Center,  >> 
 mailto:mack.joseph@xxxxxxx ph# 919-541-0007, RTP, NC, USA >>   >>  _______________________________________________ >> 
 LinuxVirtualServer.org mailing list - lvs-users@xxxxxxxxxxxxxxxxxxxxxx >> 
 Send requests to lvs-users-request@xxxxxxxxxxxxxxxxxxxxxx >> 
 or go to http://www.in-addr.de/mailman/listinfo/lvs-users _______________________________________________ LinuxVirtualServer.org mailing list - lvs-users@xxxxxxxxxxxxxxxxxxxxxx Send requests to lvs-users-request@xxxxxxxxxxxxxxxxxxxxxx or go to http://www.in-addr.de/mailman/listinfo/lvs-users . | 
| <Prev in Thread] | Current Thread | [Next in Thread> | 
|---|---|---|
| 
 | ||
| Previous by Date: | Re: newbie-problems, Arshad Khan | 
|---|---|
| Next by Date: | help ,how to configure the lvs-dr, xiongqiuhua | 
| Previous by Thread: | RE: anyone used director failover in production?, Palmer J.D.F. | 
| Next by Thread: | Re: Re: anyone used director failover in production?, Horms | 
| Indexes: | [Date] [Thread] [Top] [All Lists] |