sx2000 dbms flag question

Have a question or want to start a discussion? Post it! No Registration Necessary.  Now with pictures!

Threaded View


The docs say that when the system is restarted or reloaded (or rebooted?)
and the "dbms flag (is) off", the system will come back up with a blank
database.  I assume that I need to set the dbms flag to "on" so that I don't
wipe out the database?  I'd like to start running daily dbms checks and
weekly prog reboots, but I dont want to screw up and lose the db.

Here's the current dbms stat

DBMS info: Lab Default DB is off
DBMS info: DBMS_INITIALIZED is on
DBMS info: Journal is on
DBMS info: Debug is off
DBMS info: Cab: 5 Dim: 12 Traffic: 5
DBMS info: Check is off

I did do a dbms flag off a few months back when I loaded some new options.
But I dont know if I had to set it back on, or if it's on automatically
after the system comes back up with a blank databse.

Thanks in advance
Dave





Re: sx2000 dbms flag question


"DBMS_INITIALIZED" is the flag
Anytime you set the DBMS Flag "OFF" before rebooting, the system will come
up with a blank (default) database. Never use the DBMS Flag Off command
unless you're loading new options or performing some operation that
specifically requires it.

The DBMS Flag ON command is an invalid command.

Valid DBMS commands are:
DBMS CHECK
DBMS CLIENT
DBMS FLAG OFF
DBMS STATUS
DBMS SAVE

There is no risk of "losing" your database by performing daily DBMS checks
or weekly programmed reboots.

As always, it is very wise to keep a current "DATA SAVE" on some type of
offline media, either on RSD or TAPE or saved on some external computer
disk media as in the case of OPS_Man



Quoted text here. Click to load it







Re: sx2000 dbms flag question



Quoted text here. Click to load it

DBMS DOWNLOAD without qualifiers should (IIRC) copy the database from the
harddisc to RAM and in doing so, set DBMS_INITIALIZED to ON. (Preparing to
be shot down in flames now...)... but I think the database is still intact
on the hard disc following a DBMS FLAG OFF and then a reboot, so DBMS
DOWNLOAD should re-instate the database.


Quoted text here. Click to load it
don't
options.




Re: sx2000 dbms flag question


When you first install software and options the flag will be off. Do a data
restore and the flag is still off. Do a data save and the flag is on and
will remain that way unless you turn it off.

DBMS checks will not turn off the flag. I always run DBMS checks. Just don't
schedule them during an activity switch or they will never finish.


Quoted text here. Click to load it




Re: sx2000 dbms flag question



Quoted text here. Click to load it


Probably best not to schedule them anytime (or close to the time)
something else is happening or scheduled to happen. I know this makes
absolutely no sense, but we used to perform voice mail message wtg refresh
starting at 5 AM, the same time the scheduled DBMS check was set to kick
off. When we did this we had complaints almost every day about a small nbr
of MSG lamps being OFF when they should have been ON.  By changing the
scheduled DBMS check to start at 04:00 instead of 05:00 all of a sudden
our MSG WTG problems vanished completely.


Re: sx2000 dbms flag question



Quoted text here. Click to load it

Err.. don't you mean "DBMS save" ?

Quoted text here. Click to load it



Re: sx2000 dbms flag question


Nope I mean a data save. When you first install software, add the options
and restore a database, the flag is off. I then do the data save and it
turns on. I think it even says at that point to do a data save. Data save
not the same as DBMS Save. I am trying to remember the difference.


Quoted text here. Click to load it




Re: sx2000 dbms flag question



Quoted text here. Click to load it
Its a DBMS save that set the flag on, A data save might as well but normaly
you just issue a dbms save to set the flag

FWIR the datasave saves all tables including empty one a dbms save just
saves the data fom tables.




Quoted text here. Click to load it




Re: sx2000 dbms flag question


You know now that I think of it you are right it is DBMS Save. Sorry its
been a while.


"Ian" <spam"AT"bathfordhill.co.uk> wrote in message
Quoted text here. Click to load it




Re: sx2000 dbms flag question



Quoted text here. Click to load it

I'd be willing to put some money on it being a "DBMS save" that sets the
flag.

In a literal sense a "DBMS SAVE" copies the database forms data (but not
the forms) from dynamic ram of the plane you are addressing to the hard
drive of the same plane. By comparison, a "DATA SAVE" captures the forms
structure along with their contents, insuring that bits and bytes are
properly formed. For this reason it is possible to recover from a
corrupted file system (database corruption) with a DATA SAVE while it is
not possible to do so with a DBMS save. Compare to DBMS SAVE implying
"raw" data and "DATA SAVE" implying formatted data. The DBMS file is what
gets corrupted while the DATA file can never be. (In a data save it is
possible to have some of the forms data missing but it cannot ever be
mal-formed, ie, corrupted). When you perform a "data restore" from a
previous "data save" you will occasionally see a warning message stating
"bad data, correct via CDE" and upon closer reading it will tell you which
form and what entry failed to restore. Not so with a DBMS download because
it's raw data.

Additionally, a "DBMS SAVE"  is *not* a backup, while a "DATA SAVE"
**IS** a backup.  True, you can perform either process to external or
removable media, but you can only restore a corrupted file system from a
DATA SAVE.




Re: sx2000 dbms flag question


Quoted text here. Click to load it
Hi

well thats sort of true. But on quie a few occasions when no datasave has
been avalible after a system crash i have used a dbms save to rebuild the
system

Its simple
Load the os from rsd
apply all the options
restart
do a datasave
do a dbms download from the dbms save you have
do a datasave
do a data restore to find any database errrors

I cant say it will always work but it has for me.

Ian




Re: sx2000 dbms flag question


<spam"AT"bathfordhill.co.uk> writes:


Quoted text here. Click to load it



In such a situation I might be looking for a new service company. Of
course w/OpsMan DataSaves can be automated, but without OM whoever is
doing the MACS darn well better be performing routine backups or else be
looking for another job. Prior to clustering & OpsMan we had shown one of
the night crew in the computer room how to perform DataSaves and it became
their responsibility every Friday evening. The following Monday morning
there'd always be an envelope in our mail chute with an RSD disk in it. Of
course there was also a copy of the most recent datasave on the hard
drive, but that's not a lot of use to you when the head carves a path into
the platter. File-Redundant systems soften that blow considerably.

Quoted text here. Click to load it

Why a datasave at this point? Normally this is the point where you'd do a
data restore (if you had a data save)

Quoted text here. Click to load it


I watched a Product Support guy go through a similar exercise on an old
SX2000 "VS" one time when I had occasion to visit the Kanata operations
(for training). It wasn't a very elegant process tho he did manage to get
the system up in spite of the fact that everything done since the DBMS
save was created was lost. I recall him saying one of the keys in being
able to rebuild a system from a DBMS save was in finding out what version
of software the DBMS save was made under. Unlike a DATA save there isn't a
conversion utility for a DBMS Download.




Re: sx2000 dbms flag question



Quoted text here. Click to load it
Not sure why, as you say its the customers responsibilty.
In all cases the customers had not done any datasaves, or the media was
corrupt or they had lost them.

Quoted text here. Click to load it
ah the perfect world.

Quoted text here. Click to load it
You get a datasave of a blank database but with the options enabled.

Quoted text here. Click to load it
Quite true.

I think the "dirtyest" restore I had to do was a few months ago at a hotel,
Who had a currupt harddrive on a microlite but had lost the MOS sheet and
password, and the system software.  When contacting Mitel(UK) the sysid of
the system refered to another system in Ireland!!!! (Good old Mitel).
To get it back and restored to  new harddrive I did an old trick from SG
days, That is drop the drive from 1 inch onto its side, put it back, boot
the system, which now booted did a vol copy to rsd then with another
software disk i had, did a vol restore. and its still going today.. But I
was sweating a bit. As a freelance you do have to fly by the seats of your
pants a bit, but when it goes well it does mean the dealer keeps using you.

Ian






Re: sx2000 dbms flag question


<spam"AT"bathfordhill.co.uk> writes:


Quoted text here. Click to load it

Or never been shown how to do them ...

Quoted text here. Click to load it

Yes.. been there and done that a few times. Their new system is a LOT
better, practically flawless! Back in the days of the Black Boxes we used
to make bets w/each other how many times it would them to generate a
"correct" MOSS sheet when we did a software upgraded or added an option.
The average was 3.

Quoted text here. Click to load it

One's ability to remain focused under fire and persevere until successful
in spite of insurmountable odds is always a salient feature. I work for a
COAM customer and in 19 years and across multiple Mitel platforms have
never once had to resort to bringing in a hired gun from the local service
agency. We've always done it all, from power & grounding to CDE and MACS.
I believe ours might actually be one of the largest 48v DC-powered SX2000s
in operation today. There are 45 cans in all. 3 redundant control nodes, 8
full DSU nodes and 34 per nodes (peripheral expansion, obviously). The DC
power is also dual redundant w/backup generators. Somewhere out there on
the switchroom floor there's a couple 3300s folded into the mix for good
measure. Yes, clustered.







Re: sx2000 dbms flag question


Quoted text here. Click to load it
Hi
DC lights never took off over here. But as size goes, I had one site that
was 5 full nodes running the central site plus 4 systems remote to main
site, this is something like 9 or so 3300 plus a couple of SX2000s for
Operators.
also one Uni had 19 nodes all clustered. and Mitels largest OPS man was at
the time I was there 56 nodes being a mixture of clusters and remote sites.

Ian




Site Timeline