Re: ZFS doing insane I/O reads

> This is a 16GB Ram server running Linux Centos 5.5 64 bit. ??? Last I checked, there were no Linux ZFS ports that were actually usable. -- John Madden / Systems Engineer III Office of Technology / Ivy Tech Community College of Indiana Free Software is a matter of liberty,

Re: Cyrus Postfix LMTP tuning

x27;t remember the specifics, just that things improved when I turned them off. It might be a mis-configuration on the number of clients versus the number of allowed servers. John -- John Madden / Sr UNIX Systems Engineer Office of Technology / Ivy Tech Community College of Indiana F

Re: Cyrus Postfix LMTP tuning

LMTP under load and somewhere along the line these options fixed them. YMMV. The usual RTFineM about these applies too: http://www.postfix.org/postconf.5.html John -- John Madden / Sr UNIX Systems Engineer Office of Technology / Ivy Tech Community College of Indiana Free Software is a ma

Re: cpu and cyrus

egardless, those who need the most recent versions of these packages often have to look outside the distribution's packages. Faux pas or not, it's the truth. :) John -- John Madden / Sr UNIX Systems Engineer Office of Technology / Ivy Tech Community College of Indiana Free

Re: cpu and cyrus

eleyDB with the right version of OpenLDAP. No thanks. It's well worth your time to maintain your own compiles and even packages of Cyrus because the package maintainers can't keep up. John -- John Madden Sr UNIX Systems Engineer / Office of Technology Ivy Tech Community Colle

Re: disabling shared namespace

>> It is possible that I provide a patch. > > Yes please. This looks like a good solution. Agreed. :) -- John Madden Sr UNIX Systems Engineer / Office of Technology Ivy Tech Community College of Indiana jmad...@ivytech.edu Cyrus Home Page: http://www.cyrusimap.org/ List A

disabling shared namespace

else break? John -- John Madden Sr UNIX Systems Engineer / Office of Technology Ivy Tech Community College of Indiana jmad...@ivytech.edu Cyrus Home Page: http://www.cyrusimap.org/ List Archives/Info: http://lists.andrew.cmu.edu/pipermail/info-cyrus/

Re: high-availability Cyrus (i.e. glusterfs)?

plain DRDB+ext4 if you don't have real shared storage. John -- John Madden Sr UNIX Systems Engineer Ivy Tech Community College of Indiana jmad...@ivytech.edu Cyrus Home Page: http://www.cyrusimap.org/ List Archives/Info: http://lists.andrew.cmu.edu/pipermail/info-cyrus/

Re: high-availability Cyrus (i.e. glusterfs)?

a PHP session store; it'll certainly fall over with IMAP loads. John -- John Madden Sr UNIX Systems Engineer Ivy Tech Community College of Indiana jmad...@ivytech.edu Cyrus Home Page: http://www.cyrusimap.org/ List Archives/Info: http://lists.andrew.cmu.edu/pipermail/info-cyrus/

Re: competition

be alleviated with such a solution. John -- John Madden Sr UNIX Systems Engineer Ivy Tech Community College of Indiana jmad...@ivytech.edu Cyrus Home Page: http://www.cyrusimap.org/ List Archives/Info: http://lists.andrew.cmu.edu/pipermail/info-cyrus/

Re: mupdate cpu, thread timeouts

On 07/12/2010 02:01 PM, Wesley Craig wrote: > On 02 Jul 2010, at 09:29, John Madden wrote: >> I'm concerned about the listener_lock timeouts. > > The listener_lock timeout means that the thread waited around for 60 > seconds to see if a connection was going to arrive. Sinc

Re: mupdate cpu, thread timeouts

resting. Can I do anything with the prefork parameter for mupdate to spread things out on more cpu's or increase concurrency? >> Also during this time, mailbox changes (CREATE/etc) >> are delayed or timeout. > > That's normal, as the mupdate master blocks changes whil

mupdate cpu, thread timeouts

ot of matching thread timeout errors on the master, also consuming cpu. Also during this time, mailbox changes (CREATE/etc) are delayed or timeout. This is a 2.3.15+murder system with about 2.8mil mailboxes, two frontends, 6 backends, and a single master. Any suggestions? John -- John Madd

Re: High Availability approaches for Cyrus

d SATA storage partitions plus a single frontend absolutely rocks for our over 450,000 users (2.6m mailboxes). We don't do HA but Murder makes it easy to do if needed. John -- John Madden Sr UNIX Systems Engineer Ivy Tech Community College of Indiana jmad...@ivytech.edu Cy

Re: Backup strategy for large mailbox stores

Out of curiousity, how good is zfs with full fs scans when running in the 100-million file count range? What do you see in terms of aggregate MB/s throughput? -- John madden Sr UNIX Systems Engineer Ivy Tech Community College of Indiana jmad...@ivytech.edu On Feb 15, 2010, at 15:43

Re: Backup strategy for large mailbox stores

NBU admin > so I dunno why this is more efficient but it worked great. Yeah, we do a stream per disk. Since the disk is pegged reads, no point in doing more than one stream per filesystem that I can see. John -- John Madden Sr UNIX Systems Engineer Ivy Tech Community College of Indiana

Re: Backup strategy for large mailbox stores

We did quite a bit with snapshots (LVM) to when we were experimenting with block-level backups but there's a performance problem there -- we were saturating GbE. Snapshot doesn't really buy you anything in terms of getting the data to tape. -- John Madden Sr UNIX Systems Engin

Re: ability to disable shared namespace?

John Madden wrote: >> Isn't this what "foolstupidclients" does? I think Blackberry might >> meet the criteria... > > Not really, and already enabled in my case. According to the docs, it just > converts a "LIST *" into "LIST INBOX*" and t

Re: Backup strategy for large mailbox stores

backup." After that one full backup, the only thing you ever run is incremental. This takes 2x your disk, but it's manageable. John -- John Madden Sr UNIX Systems Engineer Ivy Tech Community College of Indiana jmad...@ivytech.edu Cyrus Home Page: http://cyrusimap.web.cmu.edu/ C

RE: ability to disable shared namespace?

> Isn't this what "foolstupidclients" does? I think Blackberry might > meet the criteria... Not really, and already enabled in my case. According to the docs, it just converts a "LIST *" into "LIST INBOX*" and that isn't sufficient. John Cyrus Home Page: http://cyrusimap.web.cmu.edu/

ability to disable shared namespace?

I'm "this close" to writing an imap proxy that supports client-side SSL that will allow me to tweak the do this on the fly. Thanks, John -- John Madden Sr UNIX Systems Engineer Ivy Tech Community College of Indiana jmad...@ivytech.edu Cyrus Home Page: http:

Re: proxyd cpu usage

> 0.23 seconds on a 35MB mailboxes file. I thought I saw in one of you > other e-mails that yours was taking about one second? Yeah, .95 seconds in my case. Even with a 4-cpu box, our user load makes that intolerable, the latency causes things to back up. John -- John Madden S

Re: proxyd cpu usage

'driver' => 'imap', ), 'imap_config' => array ( 'children' => false, 'namespace' => array ( 'INBOX.' => array ( 'name' => 'INBOX.',

Re: proxyd cpu usage

> We set the following in imapd.conf: > > sharedprefix: ~ Public Folders (We don't use altnamespace.) -- John Madden Sr UNIX Systems Engineer Ivy Tech Community College of Indiana jmad...@ivytech.edu Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus

Re: proxyd cpu usage

high > load on the Cyrus frontends. Well, none, in general. But it's a 200MB mailboxes.db so I assume scanning that (to return nothing) is what takes up the cpu time. John -- John Madden Sr UNIX Systems Engineer Ivy Tech Community College of Indiana jmad...@ivytech.edu Cyrus Home

Re: proxyd cpu usage

g track of the original thread... You said there was high > cpu usage and slowdown during login. Did you track it back to these LIST > operations? Yes, sorry, Wesley Craig's response pointed me in that direction and that definitely seems to be the problem. John -- Joh

Re: proxyd cpu usage

> Do your users have access to each other's mailboxes? Is there are a large > number of results? a02 namespace * NAMESPACE (("INBOX." ".")) (("user." ".")) (("" ".")) -- John Madden Sr UNIX Systems Engineer Ivy Tec

Re: proxyd cpu usage

no foolstupidclients setting. My fix here is to spoof the NAMESPACE with Horde's "imap_config" parameters. You can manually specify which namespaces to recognize into a config file and voila, it'll only look at INBOX.*. I'm not sure we'll reall

Re: proxyd cpu usage

k into whether or not it can be eliminated. I don't suppose there's another fullstupidclients that improves responsiveness on this call, is there? John -- John Madden Sr UNIX Systems Engineer Ivy Tech Community College of Indiana jmad...@ivytech.edu Cyrus Home Page: http://cyrusi

Re: proxyd cpu usage

ount is currently 2.2 million (400k top-level) but we only have about 30% user load at this point. I had poked around with strace but didn't find anything obvious. John -- John Madden Sr UNIX Systems Engineer Ivy Tech Community College of Indiana jmad...@ivytech.edu Cyrus Home

proxyd cpu usage

icate_db: berkeley-nosync quota_db: skiplist subscription_db: skiplist mboxlist_db: skiplist Thanks, John -- John Madden Sr UNIX Systems Engineer Ivy Tech Community College of Indiana jmad...@ivytech.edu Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusim

unexpected mailbox reservations

to adjust quotas or create sub-folders (that theoretically should already be there and return "mailbox exists). The server isn't yet in use so it isn't a matter of anyone logging into it and, say, locking it over pop3. Any ideas? John -- John Madden Sr UNIX Systems Engineer Iv

Re: VMware for Cyrus?

to. To me, the benefits of running virtualized outweigh the pitfalls -- dealing with real OS installs on real hardware, dealing with multipathing and SAN (virtual disks are easy), etc. John -- John Madden Sr UNIX Systems Engineer Ivy Tech Community College of Indiana jmad...@ivytech.edu --

2.3.15, murder, skiplist aborts

ng the abort is the one responsible, that's fine, but how do we prevent this situation to begin with? John -- John Madden Sr UNIX Systems Engineer Ivy Tech Community College of Indiana jmad...@ivytech.edu Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyru

Re: Cyrus on 64-bit RHEL 5

course be better, but I find the overhead of Xen to be worthwhile. John -- John Madden Sr UNIX Systems Engineer Ivy Tech Community College of Indiana jmad...@ivytech.edu Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info: http

Re: choosing a file system

that's what > you might have been seeing. Of course you also mounted "noatime,nodiratime" > on both? Yes, we were using notail,noatime,nodiratime. John -- John Madden Sr. UNIX Systems Engineer Ivy Tech Community College of Indiana jmad...@ivytech.edu Cyrus Home Page: http:

Re: choosing a file system

at way. We did, however, also move from a single partition to 8 of them, so that obviously has some effect as well. John -- John Madden Sr. UNIX Systems Engineer Ivy Tech Community College of Indiana jmad...@ivytech.edu Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: htt

Re: choosing a file system

it isn't due to the number of files on those filesystems? File-level backups will slow down linearly as the filesystems grow, of course. I "solve" this by adding more spools (up to 8 at the moment with about 350k mailboxes) so they can be backed up in parallel. All on ext3.

Re: Archiving emails with Cyrus

hould move it there. You can then use Cyrus' built-in search mechanisms (squat) and have to change very little. John -- John Madden Sr. UNIX Systems Engineer Ivy Tech Community College of Indiana [EMAIL PROTECTED] Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: htt

Re: Planning for load balancing

> Can anyone advice what could help? Increase your connection count to something more than 400? -- John Madden Sr. UNIX Systems Engineer Ivy Tech Community College of Indiana [EMAIL PROTECTED] Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.

Re: Problems with load balancing cluster on GFS

lots of (relatively small) storage pools to build performance. John -- John Madden Sr. UNIX Systems Engineer Ivy Tech Community College of Indiana [EMAIL PROTECTED] Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: Migrate all to skiplist?

list. John -- John Madden Sr. UNIX Systems Engineer Ivy Tech Community College of Indiana [EMAIL PROTECTED] Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: LARGE single-system Cyrus installs?

st_unsafe I see most of our writes going to the spool filesystems, not so much the meta filesystem, so I'd prefer to see something where we can keep the main databases fsnyc()ing properly but allow the individual mailboxes to just rely on filesystem journaling. Is there a cacheandindexfil

Re: LARGE single-system Cyrus installs?

al, my money's still on Linux/LVM/Reiser/ext3. 250,000 mailboxes, 1,000 concurrent users, 60 million emails, 500k deliveries/day. For us, backups are the worst thing, followed by reiserfs's use of BLK, followed by the need to use a ton of disks to keep up with the i/o. John -- John

Re: LARGE single-system Cyrus installs?

to ext3 will make matters worse, but I have nothing else to go on. John -- John Madden Sr. UNIX Systems Engineer Ivy Tech Community College of Indiana [EMAIL PROTECTED] Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info:

Re: LARGE single-system Cyrus installs?

y about 300k emails/day. John -- John Madden Sr. UNIX Systems Engineer Ivy Tech Community College of Indiana [EMAIL PROTECTED] Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

"re-singleinstancestore" a partition?

ut has anyone had to do this before/is there a tool out there already that does it? Thanks, John -- John Madden Sr. UNIX Systems Engineer Ivy Tech Community College of Indiana [EMAIL PROTECTED] Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu

Re: Recomendations for a 15000 Cyrus Mailboxes

s is now pretty miserable, struggling to pull 1MB/s off of fibre channel.) How does your experience compare? John -- John Madden Sr. UNIX Systems Engineer Ivy Tech Community College of Indiana [EMAIL PROTECTED] Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.we

Re: Recomendations for a 15000 Cyrus Mailboxes

ry long fsck's, but I've seen the same out of ext3. But for a filesystem of 35 million mail files, I figure it's got to beat ext3 on performance, at least. ...But there don't seem to be any stats at this scale to support that. John -- John Madden Sr. UNIX Systems E

Re: Recomendations for a 15000 Cyrus Mailboxes

> My guess: ext3. ReiferFS has some very annoying weaknesses that may > affect you. Please specify those weaknesses. 250,000 mailboxes on reiserfs right now, always open to options. John -- John Madden Sr. UNIX Systems Engineer Ivy Tech Community College of Indiana [EMAIL PRO

Re: Slow lmtpd

ol/imap, etc. Either way, you want to separate not just on LVM, but on the physical spindles doing the work. John -- John Madden Sr. UNIX Systems Engineer Ivy Tech Community College of Indiana [EMAIL PROTECTED] Cyrus Home Page: http://cyrusimap.web.cmu.edu/ Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: Slow lmtpd

rrency in your MTA to the same value (or n-1). There's no way your disk system (or any other) is going to be able to handle 200 lmtpd's writing simultaneously. Even with our SAN, I only allow *3* lmtpd's to write concurrently. John -- John Madden Sr. UNIX Systems Engineer Ivy Tech

Re: Slow lmtpd

ith > lock contention in heavily loaded servers? Is there anything that can > be > done to tune this, other than disabling the duplicate test? FWIW, I turned off duplicatesuppression to no avail -- lmtpd still locks and writes to /var/imap/deliver.db. ...So are you sure it's really tu

Re: Cyrus HA & LB cluster summary

> - iSCSI storage ? (cheap GigaEthernet SAN) > > according to Wikipedia http://en.wikipedia.org/wiki/ISCSI , > only 1 iSCSI-client can be connected to 1 iSCSI-server (disk) > at a time...so this does not allow for a shared FS ? Yes, much like only one host can connect to an FC LUN at once. :) iS

Re: Moving the Cyrus Mailstore

ry (MTA?) services, move the data, adjust cyrus' configuration to point to the new mail store, and start everything back up. John -- John Madden Sr. UNIX Systems Engineer Ivy Tech Community College of Indiana [EMAIL PROTECTED] Cyrus Home Page: http://cyrusimap.web.

Re: High availability email server...

ot clustering. GFS could certainly be used in this case, but would be overkill. John -- John Madden Sr. UNIX Systems Engineer Ivy Tech Community College of Indiana [EMAIL PROTECTED] Cyrus Home Page: http://asg.web.cmu.edu/cyrus Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu List Archi

Re: High availability email server...

rent) SAN performs for such an application. (And yes, of course, filesystem issues affect performance as well.) John -- John Madden Sr. UNIX Systems Engineer Ivy Tech Community College of Indiana [EMAIL PROTECTED] Cyrus Home Page: http://asg.web.cmu.edu/cyrus Cyrus Wiki/

Re: Mailstore filesystem

True, but I do expect to reach this number on this machine in the next couple of years. ...And reiserfs has been just fine so far. Then again, I didn't even consider using ext3 at the time. John -- John Madden Sr. UNIX Systems Engineer Ivy Tech Community College of Indiana [EMAIL PROTECTED

Re: Mailstore filesystem

7;t be a problem any more. How big? ext3 STILL only supports 32000 directories within a directory. That gets to be quite a problem on large installs. John -- John Madden Sr. UNIX Systems Engineer Ivy Tech Community College of Indiana [EMAIL PROTECTED] Cyrus Home Pa

Re: Cyrus Patches used at FastMail.FM

> I'm in the process of porting a couple of patches from UMich into 2.3 > and then I'm going to make a release which includes a fix for an easily > exploitable buffer overflow in pop3d. Ehm? Does this affect the 2.2.x branch as well? John -- John Madden Sr. UNIX S

Re: postfix+cyrus errors

fix+deliver doesn't do this (although it causes other problems). John -- John Madden Sr. UNIX Systems Engineer Ivy Tech Community College of Indiana [EMAIL PROTECTED] Cyrus Home Page: http://asg.web.cmu.edu/cyrus Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu List Arc

Re: hot backups

with LVM snapshots alone? John -- John Madden Sr. UNIX Systems Engineer Ivy Tech Community College of Indiana [EMAIL PROTECTED] Cyrus Home Page: http://asg.web.cmu.edu/cyrus Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

deliver and multiple bounces for a single failure

s bounces? (Postfix 2.1.5, Cyrus 2.2.12, Debian 3.1) Thanks, John -- John Madden Sr. UNIX Systems Engineer Ivy Tech Community College of Indiana [EMAIL PROTECTED] Cyrus Home Page: http://asg.web.cmu.edu/cyrus Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: lmtpd dropping mail in wrong mailbox

(We're talking postfix-2.1.5 with "lmtp_cache_connection = no", cyrus 2.2.12) How can this be? Is there any way to find out what lmtpd's doing with the transaction? Thanks, John -- John Madden Sr. UNIX Systems Engineer Ivy Tech Community College of Indiana [EMAIL PROTECTED] Cyrus Home Page: http://asg.web.cmu.edu/cyrus Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: Backup using lvm snapshot on linux

good thing :) As a user of lvm snapshots for an install of 110k mailboxes, I can tell you that they work, on reiserfs and all. =) John -- John Madden Sr. UNIX Systems Engineer Ivy Tech Community College of Indiana [EMAIL PROTECTED] Cyrus Home Page: http://asg.web.cmu

Re: lmtpd dropping mail in wrong mailbox

s. It's skiplist. John -- John Madden Sr. UNIX Systems Engineer Ivy Tech Community College of Indiana [EMAIL PROTECTED] Cyrus Home Page: http://asg.web.cmu.edu/cyrus Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: lmtpd dropping mail in wrong mailbox

the moment, I'm actually investigating what another subscriber mentioned -- loss of sync on the socket as Postfix sends the message. To resolve this, I've turned off its lmtp connection cache. I won't know if this is actually a sufficient fix for probably another week or so since th

Re: lmtpd dropping mail in wrong mailbox

0 lmtp_over_quota_perm_failure: 1 autocreatequota: 15360 autocreateinboxfolders: Sent|Drafts|Trash autosubscribeinboxfolders: Sent|Drafts|Trash autosubscribesharedfolders: user.PublicFolders | user.PublicFolders.SPAM Can | user.PublicFolders.Ham Bone fulldirhash: 1 hashimapspool: 1 mailnotifier: mailto n

lmtpd dropping mail in wrong mailbox

ED]" is. Could this be why some main is randomly [?] delivered incorrectly? (I've got the requisite info in saslauthd.conf so that authentication works properly. No problems there.) Thanks, John -- John Madden UNIX Systems Engineer Ivy Tech Community College of Indiana [EMAIL PR

Re: odd incorrect delivery problem

ource_search_base = ou=People,dc=ivytech,dc=edu ldapsource_query_filter = (&(o=Mail)(|(uid=%u)(mailLocalAddress=%u))) ldapsource_debuglevel = 0 ldapsource_result_attribute = uid ldapsource_bind = no -- John Madden UNIX Systems Engineer Ivy Tech Community College of Indiana [EMAIL PROTECTED] Cyrus Home Page

odd incorrect delivery problem

ap/sieve And because we're talking LDAP here, saslauthd.conf: ldap_servers: ldap://ldap.ivytech.edu ldap_search_base: ou=People,dc=ivytech,dc=edu ldap_auth_method: bind ldap_port: 389 ldap_version: 3 ldap_verbose: on ldap_debug: 10 ldap_filter: (&(uid=%u)(o=Mail)) Thanks, John -- John M

Re: Creating INBOXes noninteractively

> The autocreatequota option is a possibility, but there would be fewer > support calls if we could create the inboxes for them. Look for the autocreate patches for mailboxes. We use it for creation of mailboxes and auto-subscription to public folders, works great. John -- John

Re: Disabling client IP reverse DNS?

;s what you're hoping to avoid. John -- John Madden UNIX Systems Engineer Ivy Tech Community College of Indiana [EMAIL PROTECTED] Cyrus Home Page: http://asg.web.cmu.edu/cyrus Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: improving concurrency/performance

k-based and that I'm just going to have to deal with it from that angle. John -- John Madden UNIX Systems Engineer Ivy Tech Community College of Indiana [EMAIL PROTECTED] Cyrus Home Page: http://asg.web.cmu.edu/cyrus Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu List Arc

Re: improving concurrency/performance (fwd)

being spent in fdatasync >> and fsync. Actually, the thread just got off topic quickly -- I'm running this on reiserfs, not ext3. ...And I've got it mounted with data=writeback, too. But thanks for the info, Andrew. John -- John Madden UNIX Systems Engineer Ivy Tech Commu

Re: improving concurrency/performance

> Hm. I'd definitely take a second look at your ds6800 configuration ... How is > your > write cache configured there? Let's just say they're not terribly clear on that. :) -- John Madden UNIX Systems Engineer Ivy Tech Community College of Indiana [EMAIL PROTECTE

Re: improving concurrency/performance

errors column for the open() call on this strace: % time seconds usecs/call callserrors syscall -- --- --- - - 1.070.019902 17 622130 open Why 130 errors? I assume if there's an error that the call is re-trie

Re: improving concurrency/performance

BOX, pulls out the first message and logs out. I'm able to do about 230 of those per second, so at least the read performance is more than acceptable. (And the client box here, a 4-CPU Opteron 850, is definitely the bottleneck anyway.) John -- John Madden UNIX Systems Engine

Re: improving concurrency/performance

7;t ever pegging them out -- nothing ever goes into iowait. The bottleneck is elsewhere... John -- John Madden UNIX Systems Engineer Ivy Tech Community College of Indiana [EMAIL PROTECTED] Cyrus Home Page: http://asg.web.cmu.edu/cyrus Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu

improving concurrency/performance

), so I'm at a loss. Is there a general checklist of things to have a look at? Are their tools to look at the metrics of the skiplist db's (such as Berkeley's db_stat)? Am I doomed to suffer sub-par performance as long as IMAP writes are happening? Migration's coming on the 24th.

Re: High-Availability IMAP server

d pray a bunch that split-brain doesn't happen. :) John -- John Madden UNIX Systems Engineer Ivy Tech Community College of Indiana [EMAIL PROTECTED] Cyrus Home Page: http://asg.web.cmu.edu/cyrus Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu List Archives/Info: http://asg.web.cm

RE: Tutorial on how to backup/restore cyrus

nd that's the big question: How inconsistent can things be without shutting down cyrus? John -- John Madden UNIX Systems Engineer Ivy Tech Community College of Indiana [EMAIL PROTECTED] Cyrus Home Page: http://asg.web.cmu.edu/cyrus Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu Li

Re: Is there a limit to number of mailboxes in cyrus

>> FWIW, I've experimented with 750k mailboxes on a single system with 8GB RAM >> and >> we >> plan to put that number in production in a couple of months here. > > Ouch, 750k? How many concurrent accesses? Most likely less than a thousand. John -- Jo

Re: Is there a limit to number of mailboxes in cyrus

ed between all imapd processes. John -- John Madden UNIX Systems Engineer Ivy Tech Community College of Indiana [EMAIL PROTECTED] Cyrus Home Page: http://asg.web.cmu.edu/cyrus Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: Is there a limit to number of mailboxes in cyrus

ated. FWIW, I've experimented with 750k mailboxes on a single system with 8GB RAM and we plan to put that number in production in a couple of months here. John -- John Madden UNIX Systems Engineer Ivy Tech Community College of Indiana [EMAIL PROTECTED] Cyrus Home Page: http://asg.web.cm

Re: imap transfer

l be sending them new passwords that as of such-and-such a date will be what they use to log in. (Generate this list beforehand, then drop the new passwords into place on the migration date.) John -- John Madden UNIX Systems Engineer Ivy Tech Community College of Indiana [EMAIL PROTECT

Re: flat mailboxes performances for around 1500 entries

le-instance housing 750k accounts and (at minimum) 3.7 million mailboxes on a single SAN partition. Not flat files, though. :) John -- John Madden UNIX Systems Engineer Ivy Tech Community College of Indiana [EMAIL PROTECTED] Cyrus Home Page: http://asg.web.cmu.edu/cyrus C

RE: imap transfer

> Ah yeah this is good, too. I didn't think of that in my response. This way > you don't need to distribute new passwords. Yeah, sorry about that. In our situation, we're changing the password hash, so we've got to change them. John -- John Madden UNIX Systems E

Re: moving message on nearly full mailbox with IMAP

. So what *is* the solution? -- John Madden UNIX Systems Engineer Ivy Tech Community College of Indiana [EMAIL PROTECTED] --- Cyrus Home Page: http://asg.web.cmu.edu/cyrus Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: moving message on nearly full mailbox with IMAP

I'm sure we've all had problems with this. Is it wise to modify these clients to instead FETCH, delete/expunge, then STORE? John -- John Madden UNIX Systems Engineer Ivy Tech Community College of Indiana [EMAIL PROTECTED] --- Cyrus Home Page: http://asg.web.cmu.edu/cyrus Cyrus

Re: backup without stopping the imap server?

> I'm using LVM snapshot on linux box and it work perfectly But a filesystem-level snapshot isn't a clear copy of what's uncommitted to the DB's. I still haven't heard how bad a situation it is if the db's in the 'db' directory are corrupted -- what

Re: backup without stopping the imap server?

al bdb methods, so what about the skiplist db's? Isn't everything else just text files and the per-mailbox cache/etc db's that are rebuilt all the time anyway? John -- John Madden UNIX Systems Engineer Ivy Tech Community College of Indiana [EMAIL PROTECTED]

Re: 64bit Cyrus?

though I've got no comparable 32-bit-only boxes to do any performance comparisons on. John -- John Madden UNIX Systems Engineer Ivy Tech State College [EMAIL PROTECTED] --- Cyrus Home Page: http://asg.web.cmu.edu/cyrus Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu List Archives/

Re: Redundant Imap Servers

ny other application: two servers + shared storage + HA. John -- John Madden UNIX Systems Engineer Ivy Tech State College [EMAIL PROTECTED] --- Cyrus Home Page: http://asg.web.cmu.edu/cyrus Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

Re: murder or not murder

> I need to create some big nice mail architecture, which should be able to > grow nicely. > So, I have a big nfs nas, and, I have these "small" xeon 2.8 with 1G of ram. See the docs -- Cyrus + NFS = No. -- John Madden UNIX Systems Engineer Ivy Tech State Colle

Re: MURDER : tuning for increasing load

I assume reading berkeley's docs on DB_CONFIG would be a good place to start. John -- John Madden UNIX Systems Engineer Ivy Tech State College [EMAIL PROTECTED] --- Cyrus Home Page: http://asg.web.cmu.edu/cyrus Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu List Archives/Info: http://asg.web.

Re: MURDER : tuning for increasing load

y box would be ok (and management more easily swallowed "4 CPU box with 8 GB RAM" than they did "lots of little boxes.") John -- John Madden UNIX Systems Engineer Ivy Tech State College [EMAIL PROTECTED] --- Cyrus Home Page: http://asg.web.cmu.edu/cyrus Cyrus Wiki/FAQ: ht

Re: best filesystem for imap server

> I think they use capacitors that will hold enough charge to allow > flushing the buffers to disk when there's a power loss. And another set of caps to keep the spindles spinning so that data can be written? I'm not yet willing to buy the bridge you're selling. :) John

Re: best filesystem for imap server

want to discuss this on postfix-users, as you may have something bad in your config. John -- John Madden UNIX Systems Engineer Ivy Tech State College [EMAIL PROTECTED] --- Cyrus Home Page: http://asg.web.cmu.edu/cyrus Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu List Archives/Info: http

Re: best filesystem for imap server

> Is this strictly referencing UFS on Solaris? Or is this also true with > UFS on *BSD where UFS_DIRHASH is present? I was, yes, but I have no experience with it on BSD. "DIRHASH" sure sounds nice. :) John -- John Madden UNIX Systems Engineer Ivy Tech State College

  1   2   >