Two logical logs marked as Current ?
Has anyone seen such a problem before - 2 logs marked as 'C' i.e. current ?
See the bottom of the email for the online.log. This all resulted in a full
system restore, Informix Tech Support didn't give an information and were
very defensive about bugs in the server ? All attempts to free up the
logical logs failed. I would appreciate an insight anyone has to offer -
without an identification of what happened who knows whether it will happen
again !
uname -a :- SunOS 5.6 Generic sun4u sparc SUNW,Ultra-4
INFORMIX-Universal Server Version 9.14.UC2 -- Fast Recovery (CKPT REQ) --
Up 00:09:01 -- 229928 Kbytes
Blocked:CKPT
Physical Logging
Buffer bufused bufsize numpages numwrits pages/io
P-2 7 16 0 0 0.00
phybegin physize phypos phyused %used
10003f 15000 5117 23 0.15
Logical Logging
Buffer bufused bufsize numrecs numpages numwrits recs/pages pages/io
L-1 4 16 0 0 0 0.0 0.0
Subsystem numrecs Log Space used
Buffer Waiting
Buffer ioproc flags
L-1 10c038a4 0x1 0
address number flags uniqid begin size used %used
a5277b4 4 U-B---- 130 1200035 50000 50000 100.00
a5277d0 5 U-B---- 131 120c385 50000 50000 100.00
a5277ec 6 U-B---- 132 12186d5 50000 50000 100.00
a527808 7 U-B---- 133 1224a25 50000 50000 100.00
a527824 8 U-B---- 134 1230d75 50000 50000 100.00
a527840 9 U-B---- 135 123d0c5 50000 50000 100.00
a52785c 10 U-B---- 136 1249415 50000 50000 100.00
a527878 11 U-B---- 137 1255765 50000 50000 100.00
a527894 12 U-B---- 138 1261ab5 50000 49334 98.67
a5278b0 13 U---C-- 139 126de05 50000 170 0.34
a5278cc 14 F------ 0 127a155 50000 0 0.00
a5278e8 15 F------ 0 12864a5 50000 0 0.00
a527904 16 F------ 0 12927f5 50000 0 0.00
a527920 17 F------ 0 129eb45 50000 0 0.00
a52793c 18 F------ 0 12aae95 50000 0 0.00
a527958 19 F------ 0 12b71e5 50000 0 0.00
a527974 20 F------ 0 12c3535 50000 0 0.00
a527990 21 U---C-L 129 12cf885 50000 50000 100.00
13:54:28 Checkpoint Completed: duration was 1 seconds.
13:56:09 Logical Log 128 Complete.
13:56:12 Logical Log 128 - Backup Started
13:57:34 Logical Log 128 - Backup Completed
13:59:29 Checkpoint Completed: duration was 2 seconds.
14:04:33 Checkpoint Completed: duration was 3 seconds.
14:07:08 Logical Log Files are Full -- Backup is Needed
This happen to one of our systems in a 7.x release. It was identified as a bug by informix. Geoffrey Poole <Geoffrey.Poole@ogs.co.uk> wrote: > >Has anyone seen such a problem before - 2 logs marked as 'C' i.e. current ? >See the bottom of the email for the online.log. This all resulted in a full >system restore, Informix Tech Support didn't give an information and were >very defensive about bugs in the server ? All attempts to free up the >logical logs failed. I would appreciate an insight anyone has to offer - >without an identification of what happened who knows whether it will happen >again ! > > -- Melvin Mariney e-mail: melvin.mariney2@bridge.bellsouth.com
Hi!
Geoffrey Poole wrote:
> Has anyone seen such a problem before - 2 logs marked as 'C' i.e. current ?
Yes, we have!
9.1x is suffering a bug which has been fixed in 7.x. That is, when it is
in fast recovery and at the same time you try to use ontape it gets to
this kinda wierd state where there is two current logical logs.
This means that whenever you use ontape you have to check first what
state the database is in!!!
Here's a example script:
#!/bin/sh
# Added sniffing of IUS state before ontape:
#
# 5 which is online
# 255 is offline
# 0 is initialization
# 1 is Quiescent
# 2 is Fast Recovery
onstat - > /dev/null
STATE=$?
if [ $STATE = "5" ] ; then echo "\\ny\\n"|ontape -a 2>&1 >>
/var/informix/ontape_a_status ; else echo "Online in wrong state $STATE"
>> /var
/informix/ontape_a_status ; fi
exit 0
Informix support told us that this problem is not going to be fixed
until version 9.2.
Cheers,
Andreas