[Fwd: Re: serious over quota problem]
The cc: didn't seem to work so here it goes. Original Message Subject: Re: serious over quota problem Date: Wed, 11 Feb 2004 11:19:57 -0800 From: Joao Pedras [EMAIL PROTECTED] To: David R Bosso [EMAIL PROTECTED] CC: [EMAIL PROTECTED] References: [EMAIL PROTECTED] [EMAIL PROTECTED] [EMAIL PROTECTED] [EMAIL PROTECTED] So one possible solution for the problem would be to create separate quotas for the sub-folders, instead of dealing with one massive quota root. Would this be correct? David R Bosso wrote: See: https://bugzilla.andrew.cmu.edu/show_bug.cgi?id=1212 Cyrus IMAP uses a 32bit int to store quota in bytes internally, this causes problems with large quotas. While the bug says 2GB, from what I saw in the code a while back it's an unsigned long which should max out at 4GB. Can one of the developers clarify 2/4GB as the maximum? -David --On Tuesday, February 10, 2004 4:36 PM -0800 Joao Pedras [EMAIL PROTECTED] wrote: Hello Ken, Ken Murchison wrote: Joao Pedras wrote: Hi all! I have this user's box which is NOT over quota but Cyrus insists on saying that it is to sendmail. Also I am not able to drag messages into it. I have ran 'reconstruct' on the the folder with '-r -f' and 'cyrquota -f' also. I have increased/decreased the quota... The problem does not go away. Could someone please provide some insight on this issue? How do you know for a fact that the user is not over quota? Well... the a 'du' on the filesystem shows the usage around 2Gb. The quota root was set to 5,000,000. Plus 'cyrquota' would report around 40% usage which agrees with the number I just gave you. Keep in mind that a quotaroot includes ALL submailboxes. And that is the reason why I check 'du' on filesystem folder. Are there \Deleted messages which have not been expunged? Not quite sure. Wouldn't these count? Is the MUA using a Trash folder? In the user's inbox? Yes, but that one is empty. Wouldn't that also add up to the quota root? Two additional details: 1) it's cyrus 2.0.17 2) I attempted to set the quota to 50,000,000 and the problem went away. With that quota value the usage is down to 4%. Thanks Ken!!! --- Home Page: http://asg.web.cmu.edu/cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html --- Home Page: http://asg.web.cmu.edu/cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
serious over quota problem
Hi all! I have this user's box which is NOT over quota but Cyrus insists on saying that it is to sendmail. Also I am not able to drag messages into it. I have ran 'reconstruct' on the the folder with '-r -f' and 'cyrquota -f' also. I have increased/decreased the quota... The problem does not go away. Could someone please provide some insight on this issue? Thanks. :) --- Home Page: http://asg.web.cmu.edu/cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
Re: serious over quota problem
Hello Ken, Ken Murchison wrote: Joao Pedras wrote: Hi all! I have this user's box which is NOT over quota but Cyrus insists on saying that it is to sendmail. Also I am not able to drag messages into it. I have ran 'reconstruct' on the the folder with '-r -f' and 'cyrquota -f' also. I have increased/decreased the quota... The problem does not go away. Could someone please provide some insight on this issue? How do you know for a fact that the user is not over quota? Well... the a 'du' on the filesystem shows the usage around 2Gb. The quota root was set to 5,000,000. Plus 'cyrquota' would report around 40% usage which agrees with the number I just gave you. Keep in mind that a quotaroot includes ALL submailboxes. And that is the reason why I check 'du' on filesystem folder. Are there \Deleted messages which have not been expunged? Not quite sure. Wouldn't these count? Is the MUA using a Trash folder? In the user's inbox? Yes, but that one is empty. Wouldn't that also add up to the quota root? Two additional details: 1) it's cyrus 2.0.17 2) I attempted to set the quota to 50,000,000 and the problem went away. With that quota value the usage is down to 4%. Thanks Ken!!! --- Home Page: http://asg.web.cmu.edu/cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
Re: some error messages
James, I am using sendmail. Further investigation showed that sendmail's queue was huge due to an over quota account which gets a lot of mail. Every half hour sendmail with do its thing and the system would just go to an extreme high load. Sendmail would refuse connections for about a minute while the load came down. The part that concerns me about Cyrus is that the above mentioned mail box was not over quota. Nevertheless sendmail was not able to deliver a single message there. I am not sure if the message config/deliverdb/deliver-t.db: DB_INCOMPLETE: Cache flush was unable to complete is associated with the mail box quota problem or not. Thanks! James A. Pattie wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Joao Pedras wrote: | Hi all! | | I wonder if someone could help me with the following error messages: | | 1) imap/socket/lmtp: Connection refused by localhost | | The socket is there as it has always been. It just in the last few days. :| what mta are you using? If sendmail, this has happened to me and it turned out to be the hoststat was getting a connection refused/timedout error and then having to wait 30 minutes before it would try again. In the meantime it was just immediately saying the connection was down and queuing mails. I found by killing all sendmail processes, restarting cyrus and sendmail and then purging the hoststat database will fix this issue, until it happens again. I do not know why this happens and if anyone else has input/insight I too would be interested. :) - -- James A. Pattie [EMAIL PROTECTED] Linux SysAdmin / Systems Programmer -BEGIN PGP SIGNATURE- Version: GnuPG v1.2.3 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iD8DBQFAH8vxtUXjwPIRLVERAtNKAKCnH9/p66h28USGDp+naOgEg6rNPQCeNmJb EdHdCtPu8BZllzEup/2Dfyw= =9RaU -END PGP SIGNATURE-
some error messages
Hi all! I wonder if someone could help me with the following error messages: 1) imap/socket/lmtp: Connection refused by localhost The socket is there as it has always been. It just in the last few days. :| Also in the last few days: 2) config/deliverdb/deliver-t.db: DB_INCOMPLETE: Cache flush was unable to complete I would apreciate your help. Thanks
need help with noselect
Greetings! I am running the following version of cyrus: name : Cyrus version: v2.0.17 vendor : Project Cyrus support-url: http://asg.web.cmu.edu/cyrus os : FreeBSD os-version : 4.9-RC environment: Cyrus SASL 1.5.27 Sleepycat Software: Berkeley DB 3.3.11: (July 12, 2001) OpenSSL 0.9.7c 30 Sep 2003 TCP Wrappers A non-user box (general) which is acessible to several users became unresponsive (mozilla mail, outlook). When I list the mailboxes I get the following: localhost.mydomain.org lm general% general (\Noselect) general2 generalquest general1 generalerror localhost.mydomain.org 1) how can I fix this issue? 2) what might have caused it? I appreciate your help. Thanks.
Re: need help with noselect
Also let me add that sendmail can not deliver e-mail to the box. Quoting Joao Pedras [EMAIL PROTECTED]: Greetings! I am running the following version of cyrus: name : Cyrus version: v2.0.17 vendor : Project Cyrus support-url: http://asg.web.cmu.edu/cyrus os : FreeBSD os-version : 4.9-RC environment: Cyrus SASL 1.5.27 Sleepycat Software: Berkeley DB 3.3.11: (July 12, 2001) OpenSSL 0.9.7c 30 Sep 2003 TCP Wrappers A non-user box (general) which is acessible to several users became unresponsive (mozilla mail, outlook). When I list the mailboxes I get the following: localhost.mydomain.org lm general% general (\Noselect) general2 generalquest general1 generalerror localhost.mydomain.org 1) how can I fix this issue? 2) what might have caused it? I appreciate your help. Thanks.
Re: need help with noselect
Ken, thanks for you prompt reply. :) Ken Murchison wrote: Joao Pedras wrote: Also... :) A reconstruct doesn't seem to find the mailbox: %reconstruct general general: Mailbox does not exist The mailbox doesn't exist in mailboxes.db. That's bad. The folder does exit under the mail partition. I get the same error when trying to list the box or its ACL in cyradm. It exists, because it has submailboxes. Do any cyrus.* files exist in general/? They don't. My guess is either the mailbox has never existed, or was deleted (hence the \Noselect) I was afraid that was the cause of this mess. Do you have any sugestions? I was thinking of creating a new box, copy all the stuff there (messages, sub-boxes and their messages) and then run a 'reconstruct -r general'. Thanks for your help.
Re: need help with noselect
Ken Murchison wrote: Are the submailboxes missing too? They are not missing. They exist with their cyrus.* files. It is just the top one which contains nothing (except for the sub-dirs of the other mailboxes). I have been trying to repeat what the user might have done in another server but when I delete the top-level mailbox everything is removed (including sub-mailboxes). I have tried the following: - create a new mailbox - copy the directories of the other sub-folders from the damaged main box - get rid of the cyrus.* in the sub-folders - running 'reconstruct -r -f general' - voila... everything is back! The permissions of the top-level mailbox are inherited by the others, btw. 1) Does this sound like a good route or the corrupted mailbox still has hope? 2) I guess from now on I should make sure the top-level mailbox doesn't have a 'lrswipcd' acl, correct? Althought it is ok with sub-mailboxes, or not? I granted the user permissions for a few hours fo her to create some sub-mailboxes... :( Thanks Ken!!!
Re: need help with noselect
Ok... the way I fixed the problem is the way I described below, for future record. Cyrus rocks! Thanks!!! Joao Pedras wrote: Ken Murchison wrote: Are the submailboxes missing too? They are not missing. They exist with their cyrus.* files. It is just the top one which contains nothing (except for the sub-dirs of the other mailboxes). I have been trying to repeat what the user might have done in another server but when I delete the top-level mailbox everything is removed (including sub-mailboxes). I have tried the following: - create a new mailbox - copy the directories of the other sub-folders from the damaged main box - get rid of the cyrus.* in the sub-folders - running 'reconstruct -r -f general' - voila... everything is back! The permissions of the top-level mailbox are inherited by the others, btw. 1) Does this sound like a good route or the corrupted mailbox still has hope? 2) I guess from now on I should make sure the top-level mailbox doesn't have a 'lrswipcd' acl, correct? Althought it is ok with sub-mailboxes, or not? I granted the user permissions for a few hours fo her to create some sub-mailboxes... :( Thanks Ken!!!
dberrors
Hello all I am trying to run 2.0.12 on linux sparc. I keep getting the same error and I am coming to that point that I do not what else to do :( I have seen similar errors on the list, but the solutions don't seem to fit my bill. Any help would be deeply appreciated. This is what I get in imapd.log when I start master : Feb 27 00:27:19 tricana master[26917]: process started Feb 27 00:27:19 tricana master[26918]: about to exec /usr/cyrus/bin/ctl_mboxlist Feb 27 00:27:19 tricana ctl_mboxlist[26918]: running mboxlist recovery Feb 27 00:27:19 tricana ctl_mboxlist[26918]: DBERROR db3: /var/imap/db/__db.001: unable to acquire environment lock: Inval id argument Feb 27 00:27:19 tricana ctl_mboxlist[26918]: DBERROR: dbenv-open '/var/imap/db' failed: Invalid argument Feb 27 00:27:19 tricana master[26917]: process 26918 exited, status 75 Feb 27 00:27:19 tricana master[26919]: about to exec /usr/cyrus/bin/ctl_deliver Feb 27 00:27:19 tricana ctl_deliver[26919]: DBERROR: dbenv-open failed: Invalid argument Feb 27 00:27:19 tricana master[26917]: process 26919 exited, status 1 Feb 27 00:27:19 tricana master[26917]: ready for work Feb 27 00:27:19 tricana master[26920]: about to exec /usr/cyrus/bin/ctl_mboxlist Feb 27 00:27:19 tricana ctl_mboxlist[26920]: checkpointing mboxlist Feb 27 00:27:19 tricana master[26917]: process 26920 exited, status 75 Feb 27 00:27:19 tricana ctl_mboxlist[26920]: DBERROR db3: /var/imap/db/__db.001: unable to acquire environment lock: Inval id argument Feb 27 00:27:19 tricana ctl_mboxlist[26920]: DBERROR: dbenv-open '/var/imap/db' failed: Invalid argument Feb 27 00:27:19 tricana master[26921]: about to exec /usr/cyrus/bin/ctl_deliver Feb 27 00:27:19 tricana ctl_deliver[26921]: DBERROR: dbenv-open failed: Invalid argument Feb 27 00:27:19 tricana master[26917]: process 26921 exited, status 1 Tkx for your time. Joao
