Enabling ISS service

[btisdall]

Omstat is showing ISS curently stopped and I gather from elsewhere here that this is the install default. I'm now preparing to get it started and want to double-check the procedure, which from the Admin guide & the "Corruption..." STN docs I summarize thus:

Ensure ISL_DISABLE_LOGGING=TRUE is NOT set in general.cfg

Ensure ISL_LOG_IF_OFF=TRUE is NOT set in general.cfg

run $ omscan -A -a -l /path/to/file

Fix errors if necessary and re-run to ensure no errors.

run $ omupdtis -i -l /path/to/file

do $ omscan -t to determine logfiles needed to update database with changes that were logged during the omscan, apply logs.

Start ISS by issuing: $ omon -a iss <-- or do I do this at the beginning??

Make ISS startup persistent with: $ omsetsvc -a iss Y

Am I close?

Regards,

Ben Tisdall

[btisdall]

Sorry to hustle this, but feeling vulnerable here until I can get it going.

Omstat -a shows nothing in the time/date column for ISS and no refs in .../logs/fatal so I presume it never started in the first place.

Cheers.

[florian]

well, you're not really vulnerable. The ISS is needed when you want to be able to do Single User Restore from a full online backup. Otherwise it is ok to leave it turned off.

This is a small cookbook on how to proceed with SUR, in case someone needs it. more information can be found in the ISS, ISDB and SUR Scalix Technote available in the Knowledgebase.

1. Make sure ISS is running

omon iss
omsetsvc -a iss Y

2. Create initial Database "0"

omscan -aA -l /tmp/scanlog
omupdtis -I -l /tmp/scanlog

omscan -t

Last omscan tool run on 10.10.05 at 09:29:12; duration 21 minute(s).

omupdtis -r -f 10.10.05 -F 09:29:12


3. Create Backup

Do Snapshot, record time of snapshot, e.g. 10.10.05 10:30:20
Do Backup from Snapshot using Backup tool

4. If needed, perform Single User Restore

a) Create copy of Database "0"

cd /var/opt/scalix/is

cp -a 0 1

Note: replace "1" with any unused number

b) Apply log records to copy until time of backup run

omupdtis -d 1 -t 10.10.05 -T 10:30:20

Note: -d must be the same as number of copy of database from Step a)
-t and -T must have exact date and time of snapshot

c) Generate list of needed files

omprepsur -n "Mr User/scalix,demo" -d /tmp/workdir -f /tmp/filelist.txt
-i 1 -t 10.10.05 -T 10:30:20

Note: This is user specific, so name is name of user to be restored
Again, -i must specify database number assigned from step 1
Again, -t and -T must have exact date and time of snapshot
-d must be a non-existing or empty directory
-f must be a non-existing new filename

d) Retrieve all files listed in /tmp/filelist.txt and restore them under /tmp/workdir

Note: How to do this is specific to the backup tool used.

e) Verify all files have been retrieved

omprepsur -n "Mr User/scalix,demo" -d /tmp/workdir -f /tmp/filelist2.txt
-i 1 -t 10.10.05 -T 10:30:20

Note: -f has a new filelist filename

Check in filelist2.txt that all entries are either state O or state P, no longer
state N or state ?

f) Create an import file from directory tree

omdosur -n "Mr User/scalix,demo" -d /tmp/workdir -f /tmp/MrUser.store
-i 1 -t 10.10.05 -T 10:30:20

g) /tmp/MrUser.store will contain a omcpoutu/omcpinu-type file that contains the
mailbox of the user at the time of the backup. this can be imported into a shadow
account as for the import part of the omcpinu/omcpoutu cookbook.

5. Importing to a Shadow Account:

a) Check out Naming attributes

omshowu -n "Mr User/scalix,demo"

...
User Name : Mr User/CN=Mr User
MailNode : scalix,demo
...

b) Create temporary restore mailnode and user

omaddmn -m restore
omaddu -n "Mr User/restore/CN=Mr UserIA=xxx@xxx.com" -x -p scalix

Notes:
(1) First and Last Name MUST be the same
(2) Mailnode MUST NOT be the same
(3) CN and CN-TX MUST be the same
(4) IA MUST NOT be the same and does not matter if it's a valid address
(5) -x means to prevent user from being added to addressbook, -p sets temporary password

c) Import dump file into temporary account

omcpinu -f /tmp/MrUser.store -m restore

Note: DON'T forget to specify temporary mailnode using -m, otherwise data will
be imported into original mailbox, resulting in duplicate messages !!!!

6. Create a profile for Shadow Account in Outlook or Thunderbird and copy over needed messages.

7. Delete temporary shadow account and mailnode

omdelu -n "Mr User/restore"
omdelmn -m restore

[btisdall]

Thanks very much for your reply Florian.

You say I'm not vulnerable but if I needed to do a SUR from Friday's backup, ISS will have to have been setup and running before Friday's backup was made (or message store changes set to be written straight to the logs), right?

Am I right in thinking that in your cookbook one would probably insert the following step:

3a. User deletes vital mail/fs damage/...

So assuming all is currently well, I just follow steps 1-3 for now?

Sorry if these questions seem rather dumb, I did read STN #123920 prior to the OP...

Regards,

Ben.

[btisdall]

Hi Florian, whenever you could take another look at this thread I'd be most grateful.

I'm still not clear as to the exact operation of ISS, but have a nagging feeling that doesn't mean I can safely ignore it :)

Without ISS running restoring a single user's mail would require restoring /var/opt/scalix (or some entire subtree) and potentially losing new mail for other users, right?

Thanks & regards,

Ben.