Re: backup mail server
How long have you been using this setup... and have you ever had any problem... or needed to do a failover.. or a failback? Do you dedicate nics for the replication.. are they GB? Do you have any stats on the typical data rates for the replication under normal load? I've thought about using this setup as a poor mans san for a simple hot spare... but I've not known of anyone actually doing it in production.. so I'm interested to know your experiences. I'll assume you are running on hardware raid.. and doing backups... but what about using the LVM and snapshots? Much thanks you brave soul... Jared Lee wrote: - Are you using other tools like heartbeat or in the same kind ? If yes which tool ? Yes, we're using hearbeat. Heres the requisite config: /etc/ha.d/haresources: servname.host.com 100.102.248.46 datadisk::drbd0 cyrus postfix - From your drbd configuration file I can see that you are using /dev/sda6 as physical disk, is that your Cyrus partition (/var/spool/imap) ? sda6 is our "data" partitiion where we keep /var/spool/imap /var/imap/ /var/spool/mail and all of our configuration files. L -- -BEGIN GEEK CODE BLOCK- Version: 3.12 GIT/S/B d- s-:+ a- C$ UL$ P--- L+++$ E--- W+++ N++ o+ K- w O- M-- !V PS+ PE Y++ PGP++ t+ 5- X+ R* tv+ b++ DI+ D G e++>+++ h+ r>+++ z* --END GEEK CODE BLOCK--
Re: backup mail server
- Are you using other tools like heartbeat or in the same kind ? If yes which tool ? Yes, we're using hearbeat. Heres the requisite config: /etc/ha.d/haresources: servname.host.com 100.102.248.46 datadisk::drbd0 cyrus postfix - From your drbd configuration file I can see that you are using /dev/sda6 as physical disk, is that your Cyrus partition (/var/spool/imap) ? sda6 is our "data" partitiion where we keep /var/spool/imap /var/imap/ /var/spool/mail and all of our configuration files. L Thanks Marc drbd configuration file: resource drbd0 { protocol=C fsckcmd=fsck.ext2 -p -y inittimeout=60 skip-wait=yes disk { do-panic disk-size=66621523 } net { sync-rate=6M tl-size=5000 timeout=60 connect-int=10 ping-int=10 } on box1 { device=/dev/nb0 disk=/dev/sda6 address=10.0.0.1 port=7789 } on box2 { device=/dev/nb0 disk=/dev/sda6 address=10.0.0.2 port=7789 } } Boxes are connected together via serial and etho links. L On Saturday, January 11, 2003, at 05:10 PM, [EMAIL PROTECTED] wrote: On Sat, Jan 11, 2003 at 01:38:11PM -0500, Lee wrote: We use drbd (http://www.complang.tuwien.ac.at/reisner/drbd/) and linux-ha's (http://www.linux-ha.org/) heartbeat to create two-box mailstores (one active, one hotspare, continuously in-sync). Works beautifully. Are you using the drdb from CVS on a 2.4.x kernel? Could you provide details of the drdb version and OS off list? A copy/paste of your config file would also be great as example, if they do not contain too much sensitive data of course which shouldn't be the case... Thanks Marc
Re: backup mail server
Hi Lee, First thanks for your configuration file... I have a few questions now regarding your current setup: - Are you using other tools like heartbeat or in the same kind ? If yes which tool ? - From your drbd configuration file I can see that you are using /dev/sda6 as physical disk, is that your Cyrus partition (/var/spool/imap) ? Thanks Marc drbd configuration file: resource drbd0 { protocol=C fsckcmd=fsck.ext2 -p -y inittimeout=60 skip-wait=yes disk { do-panic disk-size=66621523 } net { sync-rate=6M tl-size=5000 timeout=60 connect-int=10 ping-int=10 } on box1 { device=/dev/nb0 disk=/dev/sda6 address=10.0.0.1 port=7789 } on box2 { device=/dev/nb0 disk=/dev/sda6 address=10.0.0.2 port=7789 } } Boxes are connected together via serial and etho links. L On Saturday, January 11, 2003, at 05:10 PM, [EMAIL PROTECTED] wrote: > > > > > > On Sat, Jan 11, 2003 at 01:38:11PM -0500, Lee wrote: >>> We use drbd (http://www.complang.tuwien.ac.at/reisner/drbd/) and >>> linux-ha's (http://www.linux-ha.org/) heartbeat to create two-box >>> mailstores (one active, one hotspare, continuously in-sync). Works >>> beautifully. > >> Are you using the drdb from CVS on a 2.4.x kernel? Could you provide >> details of the drdb version and OS off list? > > A copy/paste of your config file would also be great as example, if > they do > not contain too much sensitive data of course which shouldn't be the > case... > > Thanks > > Marc > >
Re: backup mail server
drbd configuration file: resource drbd0 { protocol=C fsckcmd=fsck.ext2 -p -y inittimeout=60 skip-wait=yes disk { do-panic disk-size=66621523 } net { sync-rate=6M tl-size=5000 timeout=60 connect-int=10 ping-int=10 } on box1 { device=/dev/nb0 disk=/dev/sda6 address=10.0.0.1 port=7789 } on box2 { device=/dev/nb0 disk=/dev/sda6 address=10.0.0.2 port=7789 } } Boxes are connected together via serial and etho links. L On Saturday, January 11, 2003, at 05:10 PM, [EMAIL PROTECTED] wrote: On Sat, Jan 11, 2003 at 01:38:11PM -0500, Lee wrote: We use drbd (http://www.complang.tuwien.ac.at/reisner/drbd/) and linux-ha's (http://www.linux-ha.org/) heartbeat to create two-box mailstores (one active, one hotspare, continuously in-sync). Works beautifully. Are you using the drdb from CVS on a 2.4.x kernel? Could you provide details of the drdb version and OS off list? A copy/paste of your config file would also be great as example, if they do not contain too much sensitive data of course which shouldn't be the case... Thanks Marc
Re: backup mail server
I am in the process of testing out this same setup at present under Linux and I have a couple of questions. 1) How large is your Cyrus installation (# of accounts, # of simultaneous 286 Accounts, usually around 10-15 simultaneous IMAP connections. Total spool size is 4.2 gigs, this includes stage and all user data. We've designed this system to scale to 1500 accounts per box, but havent fully tested under that load yet though. connections and IMAP spool size) 2) What DRBD protocol are you using A,B or C and over what file system (EXT2, LVM,& EXT3 ???) DRBD protocol C / Ext3 3) Are you using block replicated disks for both your IMAP spool & your IMAP directory (Mailboxes DB etc)? Yes, we are using block replication for everything directly related to services running on this system. L
ctl_cyrusdb problem
Hello, I have a little problem with the checkpoint... As you can see here from the logs: Jan 12 19:36:57 mail master[659]: [ID 392559 local6.debug] about to exec /opt/cyrimap/bin/ctl_cyrusdb Jan 12 19:36:57 mail ctl_cyrusdb[659]: [ID 702911 local6.notice] checkpointing cyrus databases Jan 12 19:36:57 mail ctl_cyrusdb[659]: [ID 854764 local6.error] DBERROR: error listing log files: Permission denied Jan 12 19:36:57 mail ctl_cyrusdb[659]: [ID 686478 local6.error] DBERROR: archive /var/cyrus/db: cyrusdb error Jan 12 19:36:57 mail ctl_cyrusdb[659]: [ID 854764 local6.error] DBERROR: error listing log files: Permission denied Jan 12 19:36:57 mail ctl_cyrusdb[659]: [ID 686478 local6.error] DBERROR: archive /var/cyrus/db: cyrusdb error Jan 12 19:36:57 mail ctl_cyrusdb[659]: [ID 127214 local6.notice] done checkpointing cyrus databases Jan 12 19:36:57 mail master[411]: [ID 310780 local6.debug] process 659 exited, status 0 I checked the whole content /var/cyrus to find any files or directories which would not beloing to user cyrus (group mail) but all files and directories belong to cyrus:mail (except the /var/cyrus/ltmp socket which belongs root but I think that's normal). Does someone have an idea what else I should check ? Thanks Regards Marc