IDS Forum

RE: HDR problem - ontape -p

Posted By: Whelan, John
Date: Monday, 27 June 2005, at 6:23 a.m.

In Response To: HDR problem - ontape -p (SAGAR THANAWALA)

Hi Sager,
See Comments below.

Regards,
John

-----Original Message-----
From: forum.subscriber@iiug.org [mailto:forum.subscriber@iiug.org]On
Behalf Of SAGAR THANAWALA
Sent: 27 June 2005 07:47
To: ids@iiug.org
Subject: HDR problem - ontape -p [5257]


Hello,

I am doing the steps for HDR on the designated Secondary Server . I have
taken a Level-0 backup from primary and am trying to do a restore on
secondary , but its NOT getting completed :-

On secondary , I have :-

(informix)/home/informix> onstat -

Informix Dynamic Server Version 7.31.FD6 -- Fast Recovery (CKPT REQ) --
Up 2
days 19:47:52 -- 662224 Kbytes Blocked:CKPT
---------------------------------------------------------------------------

(informix)/home/informix>onstat -d

Informix Dynamic Server Version 7.31.FD6 -- Fast Recovery (CKPT REQ) --
Up 2
days 18:33:55 -- 662224 Kbytes Blocked:CKPT

Dbspaces
address number flags fchunk nchunks flags owner name
c000000047e0e1c8 1 401 1 1 NP informix
rootdbs
c0000000496a6a90 2 401 2 1 NP informix
logdbs
c0000000496a6b78 3 401 3 1 NP informix
physdbs
c0000000496a6c60 4 401 4 14 NP informix
wlthdbs
c0000000496a6d48 5 401 18 4 NP informix
cometsdbs

c0000000496a6e30 6 401 22 1 NP informix
wiplotdbs

c0000000496a6f18 7 401 23 1 NP informix
ntcdatdbs

c0000000496a7000 8 401 24 1 NP informix
lecdatdbs

c0000000496a70e8 9 401 25 1 NP informix
cmtsumdbs

c0000000496a71d0 10 401 26 1 NP informix
ntcenhdbs

c0000000496a72b8 11 401 27 1 NP informix
pwlthdbs
c0000000496a73a0 12 2005 28 2 NDT informix
tempdbs1
c0000000496a7488 13 2005 30 2 NDT informix
tempdbs2
c0000000496a7570 14 2005 32 2 NDT informix
tempdbs3
c0000000496a7658 15 401 34 2 NP informix
logdbs1
c0000000496a7740 16 401 35 1 NP informix
phydbs1
16 active, 2047 maximum

Chunks
address chk/dbs offset size free bpages flags pathname
c000000047e0e2b0 1 1 0 750000 748519 PI-
/usr/informix
/data/data1
c0000000496761b8 2 2 0 175000 174947 PI-
/opt/informix
/data/data2
c0000000496762e0 3 3 750000 125000 124947 PI-
/opt/informix
/data/data1
c000000049676408 4 4 0 1016000 1015947 PI-
/opt/informix
/data/data3
c000000049676530 5 4 0 1016000 0 PI-
/opt/informix
/data/data4
c000000049676658 6 4 0 1016000 0 PI-
/opt/informix
/data/data5
c000000049676780 7 4 0 1016000 0 PI-
/opt/informix
/data/data6
c0000000496768a8 8 4 0 1016000 0 PI-
/opt/informix
/data/data7
c0000000496769d0 9 4 0 1016000 0 PI-
/opt/informix
/data/data8
c000000049676af8 10 4 0 1016000 0 PI-
/opt/informix
/data/data9
c000000049676c20 11 4 0 1016000 0 PI-
/opt/informix
/data/data10
c000000049676d48 12 4 0 1016000 0 PI-
/opt/informix
/data/data11
c000000049676e70 13 4 0 1016000 0 PI-
/opt/informix
/data/data12
c000000049676f98 14 4 0 1016000 0 PI-
/opt/informix
/data/data13
c0000000496770c0 15 4 0 1016000 1015997 PI-
/opt/informix
/data/data14
c0000000496771e8 16 4 0 1016000 1015997 PI-
/opt/informix
/data/data15
c000000049677310 17 4 0 1016000 1015997 PI-
/opt/informix
/data/data16
c000000049677438 18 5 0 1016000 13 PI-
/opt/informix
/data/data17
c000000049677560 19 5 0 1016000 7 PI-
/opt/informix
/data/data18
c000000049677688 20 5 0 1016000 848816 PI-
/opt/informix
/data/data19
c0000000496777b0 21 5 0 1016000 1015997 PI-
/opt/informix
/data/data20
c0000000496778d8 22 6 175000 508000 242993 PI-
/opt/informix
/data/data2
c000000049677a00 23 7 875000 50000 49943 PI-
/opt/informix
/data/data1
c000000049677b28 24 8 925000 50000 49943 PI-
/opt/informix
/data/data1
c000000049677c50 25 9 508000 508000 290493 PI-
/opt/informix
/data/data21
c000000049677d78 26 10 0 1016000 936843 PI-
/opt/informix
/data/data22
c000000049677ea0 27 11 0 1016000 489555 PI-
/opt/informix
/data/data23
c0000000496a6028 28 12 0 1016000 1015839 PI-
/opt/informix
/data/data24
c0000000496a6150 29 12 0 1016000 1015997 PI-
/opt/informix
/data/data25
c0000000496a6278 30 13 0 1016000 1015847 PI-
/opt/informix
/data/data26
c0000000496a63a0 31 13 0 1016000 1015997 PI-
/opt/informix
/data/data27
c0000000496a64c8 32 14 0 1016000 1015343 PI-
/opt/informix
/data/data28
c0000000496a65f0 33 14 0 1016000 1015997 PI-
/opt/informix
/data/data29
c0000000496a6718 34 15 0 175000 4947 PI-
/opt/informix
/data/data31
c0000000496a6840 35 16 175000 125000 74947 PI-
/opt/informix
/data/data31
c0000000496a6968 36 15 300000 200000 84997 PI-
/opt/informix
/data/data31
36 active, 2047 maximum

(informix)/home/informix> onstat -

Informix Dynamic Server Version 7.31.FD6 -- Fast Recovery (CKPT REQ) --
Up 2
days 19:47:52 -- 662224 Kbytes
Blocked:CKPT

18:33:00 Event alarms enabled. ALARMPROG = '/usr/informix/etc/log_full.sh'
18:33:09 DR: DRAUTO is 0 (Off)
18:33:09 Requested shared memory segment size rounded from 604KB to 608KB
18:33:10 Informix Dynamic Server Version 7.31.FD6 Software Serial
Number AC
N#
18:33:10 Informix Dynamic Server Initialized -- Shared Memory Initialized.
18:33:10 Dataskip is now OFF for all dbspaces
18:33:10 Restartable Restore has been DISABLED
18:33:10 Recovery Mode
18:33:15 Physical Restore of rootdbs, logdbs, physdbs, wlthdbs, cometsdbs,
wipl
otdbs, ntcdatdbs, lecdatdbs, cmtsumdbs, ntcenhdbs, pwlthdbs, logdbs1,
phydbs1 st
arted.

18:33:23 Checkpoint Completed: duration was 1 seconds.
18:33:23 Checkpoint loguniq 20875, logpos 0xd50018

18:33:24 Checkpoint Completed: duration was 0 seconds.
18:33:24 Checkpoint loguniq 20875, logpos 0xd50018

18:33:24 Checkpoint Completed: duration was 0 seconds.
18:33:24 Checkpoint loguniq 20875, logpos 0xd50018

22:31:08 Checkpoint Completed: duration was 1 seconds.
22:31:08 Checkpoint loguniq 20875, logpos 0xd50018
:q!
(informix)/usr/informix>

----------------------------------------------------------------------------
----

I want answers to the following questions :-

1] When I do ontape -p , it does ask me "take logical backup [Y/N] ?"

What should I do ?
I gave "N" and proceeded with the cold restore.

Is the above right action ?

Yes, that's right you do not want to perform a logical restore as you want
to completely wipe the secondary database and replace it with the primary.

2] Why doesnt the restore complete ?
I assume you did the following:
1. Ran ontape -s -L 0 on the primary
2. Ran onmode -d primary <secondary server name> on the secondary
3. Did either of the following:
- Copied across the Database backup created in step 1 from the primary
to the secondary.
- On the secondary, the TAPEDEV parameter is pointing to the database backup

created in step 1.

You should send on your onconfig file (onstat -c).

3] The level-0 backup on primary takes around 10 hours , should the restore
take more than that or how much time should the "ontape -p" run ?
10 hours seems a long time to run a level-0 backup. You also have a lot of
chunks
which may be causing it to take a long time or it may be that it is take a
long time
to write to a peripheral device. It could be your TAPEBLK, TAPESIZE values
are set
too low . How long does it take to perform a level-0 backup when TAPEDEV is
set to
/dev/null or when it is set to a file?
You may need to configure parameters such as OFF_RECOVERY_THREADS,
ON_RECOVERY_THREADS.
Have a look at the Informix archive and backup guide which can be found
starting at
http://www-306.ibm.com/software/data/informix/pubs/library/


Have you considered upgrading to IDS 9.4? The 2 GB chunk restriction no
longer applies in 9.40.
Although, I'm not sure how easy is to migrate from 7.3 to 9.4.

4] When it does complete, whats should be the status of the informix server
i.e. "onstat - " should show what output ?

In the online.log file, you should get the following message:
Physical Restore of rootdbs, logdbs, physdbs, wlthdbs, cometsdbs, wipl
otdbs, ntcdatdbs, lecdatdbs, cmtsumdbs, ntcenhdbs, pwlthdbs, logdbs1,
phydbs1 completed.

I don't currently have a system to test it out on but I think the secondary
should be in Fast Recovery mode.
You then must run onmode -d secondary <primary server name> on the
secondary.

If the logical-logs haven't rolled over on the primary server since step 1,
then the primary server
sends these logical logs directly to the secondary database server and
logical recovery occurs
automatically and the primary and secondary servers will become operational
at the end of the
onmode -d secondary step.

This may not happen in your case because it looks like level-0 backups and
physical restores take a long time
to complete.

Otherwise, if the logical logs have rolled over on the primary, then you
must have set LTAPEDEV
on the primary to a valid tape device. This is step 5 in the section
"Starting High-Availability Data
Replication for the First Time" that can be found in the Informix manuals.
You must copy the logical
logs from the primary to the secondary and then run ontape -l on the
secondary. The system will
become operation once this completes.

If LTAPEDEV is set to /dev/null and the logical logs have rolled over on the
primary between the time it
takes ontape -s -L 0 to complete on the primary and the time it takes onmode
-d secondary to complete, you
will NEVER get HDR to work.

There a few options you may wish to look at:

1. Speed up the time it takes for a tape backup / tape restore. Where is the
bottleneck? If the problem is
the write/read to a peripheral device, can TAPEDEV be set to a valid
file?

2. Set LTAPEDEV to a tape device. This may impact on performance and you
will also need to save EVERY logical
log to disk so it can take a lot capacity. However, if it is highly
likely that logical logs will roll over
before recovery from a HDR failure occurs then LTAPEDEV may need to be
set. Remember that the secondary server
will need to be able to access any backed up logical logs.

3. Increase the number of logical logs so that it takes longer for roll over
to occur. The number of logical
logs must be the same on both systems. Run onstat -l to see how may
logical logs you currently have
on your system. You should do some performance tests on your system to
determine how long it takes for the
logical logs to roll over in a worst case scenario i.e. when your system
is at it busiest. You should then
ensure that starting HDR for the first time takes longer then it does for
your logical logs to roll over. You
should also try starting HDR during a quiet period.


PLEASE give the answers pointwise .

Thanks in advance for ur help/suggestions ...

Regards,
Sagar

Messages In This Thread

• HDR problem - ontape -p
  SAGAR THANAWALA -- Monday, 27 June 2005, at 2:47 a.m.
  • RE: HDR problem - ontape -p
    Whelan, John -- Monday, 27 June 2005, at 6:23 a.m.
  • RE: HDR problem - ontape -p
    Support Informix -- Monday, 27 June 2005, at 8:42 a.m.
    • RE: HDR problem - ontape -p
      GRUBER, CARL M -- Monday, 27 June 2005, at 10:03 a.m.
      • RE: HDR problem - ontape -p
        June C. Hunt -- Monday, 27 June 2005, at 10:53 a.m.
  • Re: HDR problem - ontape -p
    SAGAR THANAWALA -- Wednesday, 29 June 2005, at 3:44 a.m.

IDS Forum is maintained by Administrator with WebBBS 5.12.