txarli33
asked on
Exchange logs are filling the disk up
Although the backups of our exchange are running good the logs are not getting completely cleared out.
We have two servers in a cluster and lately I had a message on the backup jobs stating that it couldn't backup the storage group from the passive node, I changed the configuration of the backup job and selected the option of backing up from teh active node if the passive wasn't available, not sure if this is the reason.
Some logs are getting cleared but not all and they keep growing
I'm running out of space and need to do something quick
Anybody can help?
The backup software is Arcserve r15, the servers are virtual (VMWare), windows 2008 SP2 and exchange 2007
We have two servers in a cluster and lately I had a message on the backup jobs stating that it couldn't backup the storage group from the passive node, I changed the configuration of the backup job and selected the option of backing up from teh active node if the passive wasn't available, not sure if this is the reason.
Some logs are getting cleared but not all and they keep growing
I'm running out of space and need to do something quick
Anybody can help?
The backup software is Arcserve r15, the servers are virtual (VMWare), windows 2008 SP2 and exchange 2007
Are you doing the database level backups or message level backups?
To get more disk space, first thing you could do is,
run,
eseutil /mk "e:\storage group path\e$$.chk"
and see what is the check point: "........." shows . This will show you what is the current log and subsequent logs files are NOT yet commited in to the database.
So move the old log files to other location and run a full backup. After a successful backup these t.logs are useless, you can safely delete them without any worry.
Second thing you have to troubleshoot what is really going on. Might be some wrong configuration in the backup software.
You said you was backuping passive node but you did not mention what kind of high availability you are using like, LCR, CCR or SCR. Anyway as I said do this and you will be okay.......
Let me know if you need additional help.........
~ Anand
run,
eseutil /mk "e:\storage group path\e$$.chk"
and see what is the check point: "........." shows . This will show you what is the current log and subsequent logs files are NOT yet commited in to the database.
So move the old log files to other location and run a full backup. After a successful backup these t.logs are useless, you can safely delete them without any worry.
Second thing you have to troubleshoot what is really going on. Might be some wrong configuration in the backup software.
You said you was backuping passive node but you did not mention what kind of high availability you are using like, LCR, CCR or SCR. Anyway as I said do this and you will be okay.......
Let me know if you need additional help.........
~ Anand
ASKER
Guys
tahnks a lot for the quick response
Stephen: i run document and database level backups every day
Anand: that seems waht i need, if can clear out logs from the drive i could make some space and then investigate the root ofthe problem as you mentioned.
As i was panicking, i started a backup at the database level just to make some space as i'm getting close to nothing. I may stop it and follow teh procedure you mention. Should i run it on the active or passive server?, will it affect any user/mail happening at the time?
tahnks a lot for the quick response
Stephen: i run document and database level backups every day
Anand: that seems waht i need, if can clear out logs from the drive i could make some space and then investigate the root ofthe problem as you mentioned.
As i was panicking, i started a backup at the database level just to make some space as i'm getting close to nothing. I may stop it and follow teh procedure you mention. Should i run it on the active or passive server?, will it affect any user/mail happening at the time?
Youre bets bet is to disable the Cluster service on the Passive node
Stop the information store on the active node and let the logs replay into the database
This will take email offline while this is in process
restart the information store and take a full on-line database backup
The logs will not be deleted unless Arcserve "sees" a full succesful backup or too much mail is coming in while the database is being backed as mail cannot be written to the database while being backed up
Stop the information store on the active node and let the logs replay into the database
This will take email offline while this is in process
restart the information store and take a full on-line database backup
The logs will not be deleted unless Arcserve "sees" a full succesful backup or too much mail is coming in while the database is being backed as mail cannot be written to the database while being backed up
In this case enabling the circular logging temporarily ia also an option. (In case you don't have too many Storage Groups)
http://technet.microsoft.com/en-us/library/bb331968(EXCHG.80).aspx
Once circular logging is enabled, the logs will be purged. As soon as the logs are purged you disable circular logging and take a full backup immediately.
http://technet.microsoft.com/en-us/library/bb331968(EXCHG.80).aspx
Once circular logging is enabled, the logs will be purged. As soon as the logs are purged you disable circular logging and take a full backup immediately.
what we have a log purging problem, we typically just run a full backup with MS backup tool so we dont have to take down the DB
ASKER
I'm a bit worried about these two options as I don't want to take email offline (virgored) and (rr1968) this is a production environment. I am more inclined to follow anand.
I do apreciate your input and may go your way if anand's doesn't give me the space I need.
thanks for your response
I do apreciate your input and may go your way if anand's doesn't give me the space I need.
thanks for your response
I have done this many times. Dismounting and mounting wouldn't take more than 30 seconds.
But you are the best judge of your environment...
But you are the best judge of your environment...
ASKER
Anand, I have run the eseutil command as you indicate.
The checkpoint shows the name of a log file created 30 min ago, are you saying that i can move out of this folder all log files older than this checkpoint?.
there actually are log files there for the last two weeks, am i safe moving out old files?
The checkpoint shows the name of a log file created 30 min ago, are you saying that i can move out of this folder all log files older than this checkpoint?.
there actually are log files there for the last two weeks, am i safe moving out old files?
ASKER
By the way, it is a CCR environment
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
it solve my inmediate problem
Hi Anand,
You are safe to move the OLD log files as mentioned in the check point header to other drive.
Does it mean I can Cut and Paste those files into the new hard drive that I created ?
But Until you take a FULL backup these logs are very important. After a full backup these logs are useless so you can delete them.
Any VSS-aware backup solution can delete/truncate/purge the log automatically right ?
You are safe to move the OLD log files as mentioned in the check point header to other drive.
Does it mean I can Cut and Paste those files into the new hard drive that I created ?
But Until you take a FULL backup these logs are very important. After a full backup these logs are useless so you can delete them.
Any VSS-aware backup solution can delete/truncate/purge the log automatically right ?