Question about logical log
Mike Ng asked whether a disk crash on the logical log disk in ONLINE mode requires restoring the whole database. Jacob Salomon said the system usually goes down; with continuous log backups, restoring only rootdbs and the log dbspaces might work, otherwise a full restore. John Bejarano said a mirrored log dbspace keeps running; unmirrored needs a cold restore. Art Kagel said restore the whole engine and mirror logs and rootdbs.
Auto-generated by Claude from the posts below — may be imperfect; read the full thread.
Hi all, I have a stupid question that if the crash of disk occurs on the disk which contain the logical log during ONLINE mode. Do I need to do restoration of the whole database to recover the database? Or the databse is still keep running and I just to change the fault disk. Thank you in advance. Mike
In article <374a9fe2@news1.us.ibm.net>, "Mike Ng" <mike@email.tech-trans.com> wrote: > Hi all, > > I have a stupid question that if the crash of disk occurs on the disk > which contain the logical log during ONLINE mode. Do I need to do > restoration of the whole database to recover the database? Or the > databse is still keep running and I just to change the fault disk. Mike, Your question is not stupid. Your situation? Weeelll... In general, if you have a crash on the disk containing the logical log, the Informix system is usually down and must be restored. Your second possibility: | Or the databse is still keep running and I just to change the fault | disk sounds like an impossible situation. If you have been good about running continuous log backups, you may be able to get away with restoring only the rootdbs and the dbspaces that hold the physical and logical logs. It's worth a try, especially if you have one of those 43gb systems. If you have not been so good about continuous log backup, it's almost sure to fail. After all, there is data on the other dbspaces (not to mention extents and new dbspaces/chunks) whose corresponding log entries have not been backed up to the tape. And due to the disk crash, you can't back it it up now. So restoring only the logs will leave your environment out of sync with the logs. In this case, your only choice is to restore the whole system. Yucch! When I was an instructor at Informix I used to preach that even if you are squeezed for disk space, *ALWAYS* mirror the root dbspaces and the dbspaces containing the physical and logical logs. And always on different physical disks. This is the best protection agains the situation you are in. Good luck. Sounds like you will need it. -- +---- Jacob Salomon -- Obligatory sesquipedalian obfuscation: ---------+ | An object of igneous, sedimentary or metamorphic mineral in combined | | states of elevated linear and rotational kinetic energy acquires no | | accumulation of bryophytic vegetation -------------------------------| --== Sent via Deja.com http://www.deja.com/ ==-- ---Share what you know. Learn what you don't.---
Mike, If the dbspace that contains your logical logs is mirrored, then it should continue to run, and you can just change out the bad disk while the database's logical logs run off of the mirror. If that dbspace is NOT mirrored, or it is mirrored and both the primary and the mirror go down. You'll need to perform a cold restore, (i.e. without the database available). Hope this helps. Kind regards, John Bejarano. Mike Ng wrote in message <374a9fe2@news1.us.ibm.net>... >Hi all, > >I have a stupid question that if the crash of disk occurs on the disk which >contain the logical log during ONLINE mode. Do I need to do restoration of >the whole database to recover the database? Or the databse is still keep >running and I just to change the fault disk. > >Thank you in advance. >Mike > >
Mike Ng wrote: > > Hi all, > > I have a stupid question that if the crash of disk occurs on the disk which > contain the logical log during ONLINE mode. Do I need to do restoration of > the whole database to recover the database? Or the databse is still keep > running and I just to change the fault disk. No you are hosed and have to restore the entire engine. MIRROR YOUR logical and physical log disks as well at ROOTDBS AT A BARE MINIMUM! Art S. Kagel