IMAP connections cannot login

[jon@adtec]

Following running out of disk space on my SUSE mail server, I added an additional drive, copied over the scalix directories (65Gb) onto the new drive and created a new mount at /var/opt/scalix (for the new drive partition) to avoid configuration issues. Everything is working however I am no login able to login via IMAP (I have outlook accessing the system using IMAP and I just keep getting prompted for the password). When I review the IMAP server log I am getting this error:
WARNING IMAP Server Da(IMAP Server Pr) 10.31.11 13:48:30
[VISTA -907] Error -907 in db_vista.
User Name: <name> Mr. / imail/CN=<name>
-> ual_OpenTraceFile
<- ual_OpenTraceFile
-> ual_OpenTraceFile
<- ual_OpenTraceFile
-> ual_SendMessage
<- ual_SendMessage
-> ual_RecvReply
-> cdl_SetTnefRoute
<- cdl_SetTnefRoute
-> ual_Init
-> InitDBSArea
<- InitDBSArea
-> ul_OpenUL
<- /build/11.2.0/src/lib/ombase/vi/vi_open.c:201[4,-907]
<- /build/11.2.0/src/lib/ombase/vi/vi_open.c:262[4,-907]
<- /build/11.2.0/src/lib/ombase/dr/dr_opendir.c:556[4,-907]

I have run the dbcheck command on all the db_vista directories, and they return with out errors. I am out of ideas ... any help?

[mikethebike]

Hi Jon,

things can sometimes get a bit screwed. There is most likley a directory lock file stuck somewhere.

If you run "omshowu -m all" do you get any error?

you will need to shut down scalix and do the following:

cd ~sys
cat dir.index

you will see something like:

2685 C 111 1000
S 0 FREEBUSY 4 VISTA 2/4/vi_dir
1035838604 0 2.0
H 0 USERLIST 108 VISTA 2/6/vi_dir
1097921242 0 3.0
S 0 SYSTEM D 109 VISTA 2/3/vi_dir
1097921584 604800 3.0

check each of those directories to see if there is a lock file, called something like "vi_dir.lock"

cd to the directories listed by going to cd ~/dir/#/# and do an ls -lrt to view the contents of the directory

remove the vi_dir.lock file

You may need to run an omdiropt with a -u option after fixing:

omdiropt -a -S -d system -t s -u

if its the userlist, run like this:

omdiropt -d userlist -t h -a s/g/i

Mick

[jon@adtec]

Hi Mick

Thanks for the response. Unfortunately still the same error.
I am able to execute the 'omshowu -m all' command successfully before and after removing lock files.

I am still getting the same error message re VISTA -907

I was able to run the 'omdiropt -d userlist -t h -a s/g/i' command successfully, but when I run 'omdiropt -a -S -d system -t s -u' or 'omdiropt -a S -d system -t s -u' I end up with:
omdiropt -a S -d system -t s -u
Starting Directory Optimization.

DO NOT ATTEMPT TO ADD ANY ENTRIES TO THE DIRECTORY BEING OPTIMIZED,
WHILST THE OPTIMIZATION IS RUNNING, AS THESE ENTRIES MAY BE LOST.
IF YOU OPTIMIZING THE SYSTEM DEFAULT DIRECTORY THEN DO NOT ATTEMPT
TO ADD ANY NEW MAILBOXES TO THE SYSTEM. IF THE DIRECTORY HAS CDA
CONFIGURED THEN OMEXECCDA MUST BE RE-RUN AFTER THE OPTIMIZATION.

Fetching all entries from Directory ... Please wait
Sorting Directory entries ... Please wait
Loading entries into new Directory ... Please wait
[OM 16951] More entries found than requested limit


Problem encountered. Failed to write Directory entries to temporary working Directory: TMP.10430

now when I view the dir.index I get:
180
C 8 1000
S 0 SYSTEM D 2 VISTA 1/2/vi_dir 1199681319 604800 4.0
S 0 FREEBUSY 3 VISTA 1/3/vi_dir 1199681365 0 4.0
H 0 USERLIST 5 VISTA 1/5/vi_dir 1320079823 0 4.0

directory 1/5/ is a new directory ... the other 2 are old. here is the listing for 1/5
ls -al /var/opt/scalix/il/s/dir/1/5
total 136
drwxrwx--- 2 scalix scalix 4096 Nov 1 03:50 .
drwxrwx--- 8 scalix scalix 4096 Nov 1 03:52 ..
-rw-rw---- 1 scalix scalix 44 Nov 1 03:50 dir_1024.dat
-rw-rw---- 1 scalix scalix 8192 Nov 1 03:50 dir_128.dat
-rw-rw---- 1 scalix scalix 8192 Nov 1 03:50 dir_128.key
-rw-rw---- 1 scalix scalix 8192 Nov 1 03:50 dir_16.dat
-rw-rw---- 1 scalix scalix 8192 Nov 1 03:50 dir_16.key
-rw-rw---- 1 scalix scalix 8192 Nov 1 03:50 dir_24.key
-rw-rw---- 1 scalix scalix 12288 Nov 1 03:50 dir_256.dat
-rw-rw---- 1 scalix scalix 8192 Nov 1 03:50 dir_32.dat
-rw-rw---- 1 scalix scalix 8192 Nov 1 03:50 dir_32.key
-rw-rw---- 1 scalix scalix 8192 Nov 1 03:50 dir_4.key
-rw-rw---- 1 scalix scalix 8192 Nov 1 03:50 dir_40.key
-rw-rw---- 1 scalix scalix 7168 Nov 1 03:50 dir_512.dat
-rw-rw---- 1 scalix scalix 8192 Nov 1 03:50 dir_64.dat
-rw-rw---- 1 scalix scalix 8192 Nov 1 03:50 dir_64.key
-rw-rw---- 1 scalix scalix 8192 Nov 1 03:50 dir_8.key
-rw-rw---- 1 scalix scalix 44 Nov 1 03:50 dir_keys.dat
-rw-rw---- 1 scalix scalix 3455 Nov 1 03:50 vi_dir.dbd
-rw-rw-rw- 1 root scalix 0 Nov 1 03:50 vi_dir.lock

The other directories contain a log file ... but /5 doesn't, don't know if this is significant or not

Any ideas as to where to go from here?
thanks

[mikethebike]

I have seen this before, I just cannot remember what the exact issue was.

Have a look at the ~/sys/FreeDirSlot file. This file holds a record of empty directories in /var/opt/scalix/il/s/dir.
If there are entries in there where a directory does not exist, or there is an entry there for a directory that is not empty, that may be the issue. Remove any files in the empty unix directories (1\4, 1\1 etc)

From what I remember, you can just remove the empty unix directories, and delete the FreeDirSlot file.
If that fails, create an empty directory (one that is not in dir.index) and create the entry in FreeDirSlot file (you cannot have a <CR><LF> at the end of that file so you should do:
echo "\1\4" > FreeDirSlot

Mick

[jon@adtec]

Hi Mick

thanks for all your help. I ended up going back to the original scalix directories that I had on another drive (which still worked) and just copying the data dir (/il/s/data) to a new drive partition, removing the existing data directory (after backing up of course approx 65Gb) to a new drive and then inserting the new drive starting point as the 'data' directory which worked fine. So overcoming the full disk issue.

I still don't understand why the copy of the original scalix directory didn't work ... there is not really that much info on the vista db ... unfortunately couldn't resolve, but at least ended up with a working solution. Next challenge is to upgrade to a new machine and latest version of Scalix.

Jon