/var/imap/msg/shutdown file
What should be in the imap/msg/shutdown file (if anything) and under what conditions is it opened? We are having some problems with LMTP delivery timeouts over our internal lan, and after running a ktrace (BSDi) on lmtpd, I found the following: 5899 lmtpd NAMI "/var/imap/msg/shutdown" 5899 lmtpd RET open -1 errno 2 No such file or directory Should I worry about this? I'm just trying to remove every possible problem. Regards, M. GambleSenior Network AdministratorThe Echo Group of Companies
Re: Cyrus 2.1.11 on BSDi 4.3.1
Does anyone have any experience doing this migration (From Berkeley to Skiplist)? We have aprox. 12,000 active mailboxes - what potential issues could we run into? What impact will this have on the performance of our IMAP server? Thanks in advance, M. Gamble On Monday, February 10, 2003, at 10:03 AM, Rob Siemborski wrote: On Sun, 9 Feb 2003, Echo Online Development Team wrote: Greetings, I hope I am posting the problem to the right list, if not, please direct me to the appropriate list. We are having a serious issue with Cyrus on BSDi 4.3.1. Over time (aprox 24 hours) Cyrus begins to slow down to the point where IMAP clients begin to time out. Restarting Cyrus fixes the problem for another 24-36 hours. The one error message we suspect is "DBERROR db4: 475 lockers" I'm assuming you're using the default Berkeley DB for the mboxlist database? If so, you should try converting to a "skiplist" format database. (dump mailbox list, recompile with --with-mboxlist-db=skiplist, undump mailbox list). -Rob -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Rob Siemborski * Andrew Systems Group * Cyert Hall 207 * 412-268-7456 Research Systems Programmer * /usr/contributed Gatekeeper
Cyrus 2.1.11 on BSDi 4.3.1
Greetings, I hope I am posting the problem to the right list, if not, please direct me to the appropriate list. We are having a serious issue with Cyrus on BSDi 4.3.1. Over time (aprox 24 hours) Cyrus begins to slow down to the point where IMAP clients begin to time out. Restarting Cyrus fixes the problem for another 24-36 hours. The one error message we suspect is "DBERROR db4: 475 lockers" I've been doing some research and found that this error message is normal, however, I have never seen any postings which refer to the number of lockers being so high. Our software setup is as follows: Software: BSDi 4.3.1 Binary Cyrus imap v2.1.11 Two custom patchs to the code - 1) Applied the NetBSD patches to ctl_cyrusdb.c which remove the trailing slash from the backup dir names, as BSDi does not support trailing slashes to the rmdir call 2) Commented out /* limit_fds(RLIM_INFINITY); */ in master.c, as this call returns -1 in BSDi. Commenting it out means that Cyrus should default to the OS max. Cyrus SASL 2.1.10 db-4.1.25 (With one patch listed on the sleepycat website) Can any one provide any insight or steps we can take to debug this setup? If any of the Cyrus developers need access to a BSDi box, we are more than willing to provide a development BSDi 4.3.1 machine. Thanks in advance, M. Gamble Echo Online Administration
