Benoit Schmid | 1 Oct 12:50 2009
Picon
Picon

Restore an imap directory that has been deleted

Hello,

We are using:
store1 SUPERUSER =seguin/ imta version
Sun Java(tm) System Messaging Server 7u2-7.04 64bit (built Jul  2 2009)
libimta.so 7u2-7.04 64bit (built 15:59:12, Jul  2 2009)
Using /opt/sun/comms/messaging64/config/imta.cnf (compiled)
SunOS store1.unige.ch 5.10 Generic_138888-05 sun4u sparc
SUNW,SPARC-Enterprise

I would like to restore an imap folder that has been deleted.

The machine is running zfs.
I have followed the sun recommendations.
Therefore there are two filesystems:
# df -h | grep prima
/store/primary-idx     xT    xG   xT     x%    /store/primary-idx
/store/primary-msg     xT   xT   xT    xx%    /store/primary-msg

I have copied. from a zfs snapshot, the directory that contains
the imap folder in:
/store/primary-msg/=user/56/75/=username
# cp -rip /snapshot/oldFolder/ /store/primary-msg/=user/56/75/=username/

Then I have run:
# reconstruct -m -p partitionUsername -u username

But oldFolder was not in the list returned by the command.
Therefore it did not see the folder.

(Continue reading)

Kelly Caudill | 1 Oct 13:31 2009
Picon

Re: Restore an imap directory that has been deleted


Benoit Schmid wrote:
> Hello,
> 
> We are using:
> store1 SUPERUSER =seguin/ imta version
> Sun Java(tm) System Messaging Server 7u2-7.04 64bit (built Jul  2 2009)
> libimta.so 7u2-7.04 64bit (built 15:59:12, Jul  2 2009)
> Using /opt/sun/comms/messaging64/config/imta.cnf (compiled)
> SunOS store1.unige.ch 5.10 Generic_138888-05 sun4u sparc
> SUNW,SPARC-Enterprise
> 
> I would like to restore an imap folder that has been deleted.
> 
> The machine is running zfs.
> I have followed the sun recommendations.
> Therefore there are two filesystems:
> # df -h | grep prima
> /store/primary-idx     xT    xG   xT     x%    /store/primary-idx
> /store/primary-msg     xT   xT   xT    xx%    /store/primary-msg

So you have configutil option store.partition.primary.messagepath set as 
well as store.partition.primary.path ??  Thus you have the .msg files on 
a different file system from the store.* files.

> I have copied. from a zfs snapshot, the directory that contains
> the imap folder in:
> /store/primary-msg/=user/56/75/=username
> # cp -rip /snapshot/oldFolder/ /store/primary-msg/=user/56/75/=username/

(Continue reading)

Benoit Schmid | 1 Oct 14:41 2009
Picon
Picon

Re: Restore an imap directory that has been deleted

Hello,

Kelly Caudill wrote:
>
> Did you only copy the .msg files?   not also the store* files to the
> appropriate primary-idx directory?
>

At first I just copied the .msg files and nothing in
primary-idx directory.

Then I just mkdir on primary-idx but I did not copy the store.idx

>
> I would also check exactly what directory you used.  Each folder
> directory name starts with = and must have certain special characters
> escaped.  But I assume you put the directories back with exactly the
> same names and locations as on the snapshot, so that should not be a
> problem.
>

In the msg it had exactly the same name because I copied it.
In the idx directory I have run
mkdir '=+Mxx de +Pyy - +Site +S & +A+Ok-curis & +A+Ok-/'
But I did not copy the store.idx or any other store.*
files.
Sould a touch store.idx in '=+Mxx de +Pyy - +Site +S & +A+Ok-curis &
+A+Ok-/' be sufficient?

>
(Continue reading)

Kelly Caudill | 1 Oct 14:55 2009
Picon

Re: Restore an imap directory that has been deleted


Benoit Schmid wrote:
> Hello,
> 
> Kelly Caudill wrote:
>> Did you only copy the .msg files?   not also the store* files to the
>> appropriate primary-idx directory?
>>
> 
> At first I just copied the .msg files and nothing in
> primary-idx directory.
> 
> Then I just mkdir on primary-idx but I did not copy the store.idx
> 
>> I would also check exactly what directory you used.  Each folder
>> directory name starts with = and must have certain special characters
>> escaped.  But I assume you put the directories back with exactly the
>> same names and locations as on the snapshot, so that should not be a
>> problem.
>>
> 
> In the msg it had exactly the same name because I copied it.
> In the idx directory I have run
> mkdir '=+Mxx de +Pyy - +Site +S & +A+Ok-curis & +A+Ok-/'
> But I did not copy the store.idx or any other store.* files.
> Sould a touch store.idx in '=+Mxx de +Pyy - +Site +S & +A+Ok-curis & +A+Ok-/' be sufficient?

Just using touch will create the file and may be enough to cause 
reconstruct -m to create the folder DB entry.

(Continue reading)

Derek Diget | 2 Oct 07:47 2009

SPF DNS lookups (spfhelo, spfmailfrom, spfnone, spfrcptto)


Sun Java(tm) System Messaging Server 7.3-11.01 64bit (built Sep  1 2009)

I was reading over ${SUBJECT} at 
<http://msg.wikidoc.info/index.php?title=Spfhelo%2C_spfmailfrom%2C_spfnone%2C_spfrcptto_Channel_Options>
and have a couple of questions.

What would be a recommended "setting" on tcp_local if a site wanted to 
reject messages that "Fail" per section 2.5.4 of RFC 4408?  "spfhelo" 
and either "spfrcptto" or "spfmailfrom" where which one depends on when 
you want to reject the message?  Or do they all "do the same thing" and 
you just want one depending on when you want to do the reject during the 
SMTP transaction?  (Above assuming that the "spf_smtp_status_*" are left 
to their defaults.)

When the page talks about "spf_smtp_status_temperror" and 
"spf_smtp_status_permerror", do these correspond to the RFC 4408 section 
2.5.6 "TempError" and section 2.5.7 "PermError" respectively?

Is there any log options available to log the rejected message?  (I have 
not started testing these channel keywords so I am asking be finding out 
myself.)  If the message are logged, I would assume that using spfrcptto 
would be able to log both MAIL FROM and RCPT TO addresses whereas 
spfmailfrom would only get the MAIL FROM.  Would the log entry be a "J" 
log action code with a ? (new?) log entry modifier code?  What would get 
logged if using spfhelo?  Would it be logged under a new LOG_CONNECTION 
action code?  (Using "Understanding the MTA Log Entry Format" at

<http://wikis.sun.com/display/CommSuite/Managing+Logging#ManagingLogging-UnderstandingtheMTALogEntryFormat> 
as a reference of action and modifier codes.)
(Continue reading)

Derek Diget | 2 Oct 08:25 2009

identnone and forwardchecktag channel keywords


Sun Java(tm) System Messaging Server 7.3-11.01 64bit (built Sep  1 2009)

When I have identnone and forwardchecktag set I get the part of 
the following header

Received: from mx-tmp.wmich.edu (mx-tmp.wmich.edu** [141.218.1.43])
   by mta.REDACTED.net (Sun Java(tm) System Messaging Server 7.3-11.01
   64bit (built Sep  1 2009))

But the way I read the forwardchecktag section of "Identtcp, 
identtcplimited, identtcpnumeric, identtcpsymbolic, identnone, 
identnonelimited, identnonenumeric, identnonesymbolic, forwardchecknone, 
forwardchecktag, forwardcheckdelete Channel Options" at

<http://msg.wikidoc.info/index.php/Identtcp%2C_identtcplimited%2C_identtcpnumeric%2C_identtcpsymbolic%2C_identnone%2C_identnonelimited%2C_identnonenumeric%2C_identnonesymbolic%2C_forwardchecknone%2C_forwardchecktag%2C_forwardcheckdelete_Channel_Options> 
the double asterisk should NOT be there.  When a lookup on the DNS PTR 
record is done from the receiving MTA (mta.REDACTED.net) for 
141.218.1.43 it gets an answer of mx-tmp.wmich.edu.  Then from the same 
system a DNS A record lookup of mx-tmp.wmich.edu is done and gets an 
answer of 141.218.1.43.  So, forward and reverse match.

Is the wiki page wrong or the code?  Does a double asterisk mean that 
forward and reverse matched where a single asterisk as documented means 
it did not match?

--

-- 
***********************************************************************
Derek Diget                            Office of Information Technology
Western Michigan University - Kalamazoo  Michigan  USA - www.wmich.edu/
(Continue reading)

Kristin Hubner | 2 Oct 20:08 2009
Picon

Re: identnone and forwardchecktag channel keywords

I have a note to myself from years ago that there is some case where
two asterisks can get appended, but no discussion of which case -- not
very helpful to you :(.

So hmm, looking at the code again: it looks to me like what is  
supposed to
happen with forwardcheck is that you get one asterisk appended if the  
A lookup
on the name outright fails (the getaddrinfo call fails), whereas if  
the A
lookup returns IP addresses but none of the IP addresses "match" the  
original
client IP, then you (should) get two asterisks appended.  But if the A  
lookup
returned address(es) and some address matched, then nothing is (should  
be) appended.

Now why you're getting two asterisks appended when you believe the  
values
should match, I'm not sure -- either there's something odd with your  
DNS,
or I'm misunderstanding the code, or (I think quite possibly) it's a  
bug.
This code really hasn't gotten exercised much -- we put it in for some
customer that I think never then used it -- and indeed, there was a  
bug # 6688901 (forwardcheck in SMTP server not working correctly)  
whereby asterisks were never
appended which didn't get noticed and fixed until MS 7.0, probably  
because no
one ever used the feature.
(Continue reading)

Marc Chapados | 2 Oct 20:18 2009
Picon

orphaned-inactive mailbox

Hi,

I tried this command on my mailbox server:

mboxutil -o -t 183  

and that gave me the number of mailboxes orphaned/inactive since 183 days.  
The result: 12261 mailboxes.

On the same server, I tried the same command but for 365 days.

mboxutil -o -t 365
The result: 10664 mailboxes.

How can we explain the result?  
For 365 days, I must have at least the same numbers of mailboxes or higher than the result for 183 days...

Any thought???

Marc Chapados | 2 Oct 21:00 2009
Picon

Re : orphaned-inactive mailbox

Forget my question...    :)

Regards!

----- Message d'origine -----
De: Marc Chapados <marc.chapados <at> telusquebec.net>
Date: Vendredi, Octobre 2, 2009 2:26 pm
Objet: [Info-iMS] orphaned-inactive mailbox
À: "Info-iMS@..." <Info-iMS@...>


> Hi,
>  
>  I tried this command on my mailbox server:
>  
>  mboxutil -o -t 183  
>  
>  and that gave me the number of mailboxes orphaned/inactive since 183 
> days.  
>  The result: 12261 mailboxes.
>  
>  On the same server, I tried the same command but for 365 days.
>  
>  mboxutil -o -t 365
>  The result: 10664 mailboxes.
>  
>  How can we explain the result?  
>  For 365 days, I must have at least the same numbers of mailboxes or 
> higher than the result for 183 days...
>  
(Continue reading)

Kristin Hubner | 2 Oct 21:18 2009
Picon

Re: SPF DNS lookups (spfhelo, spfmailfrom, spfnone, spfrcptto)


On Oct 1, 2009, at 10:47 PM, Derek Diget wrote:

>
> Sun Java(tm) System Messaging Server 7.3-11.01 64bit (built Sep  1  
> 2009)
>
> I was reading over ${SUBJECT} at
<http://msg.wikidoc.info/index.php?title=Spfhelo%2C_spfmailfrom%2C_spfnone%2C_spfrcptto_Channel_Options 
> >
> and have a couple of questions.
>
> What would be a recommended "setting" on tcp_local if a site wanted  
> to reject messages that "Fail" per section 2.5.4 of RFC 4408?   
> "spfhelo" and either "spfrcptto" or "spfmailfrom" where which one  
> depends on when you want to reject the message?  Or do they all "do  
> the same thing" and you just want one depending on when you want to  
> do the reject during the SMTP transaction?  (Above assuming that the  
> "spf_smtp_status_*" are left to their defaults.)

What domain do you want to check: the domain that the SMTP client  
claims as its hostname
on its HELO/EHLO line (spfhelo), or the domain name in the envelope  
 From address on the
MAIL FROM line (spfmailfrom or spfrcptto)?  Or both?  Note that the  
HELO claimed name
and the MAIL FROM domain name need not be the same, so these are  
potentially different
and independent checks.

(Continue reading)


Gmane