LinuxQuestions.org

LinuxQuestions.org (/questions/)
-   Linux - Server (https://www.linuxquestions.org/questions/linux-server-73/)
-   -   AMANDA Issue - Recording Takes Longer Than It Should (https://www.linuxquestions.org/questions/linux-server-73/amanda-issue-recording-takes-longer-than-it-should-923320/)

robert.wolfe 01-11-2012 03:44 PM

AMANDA Issue - Recording Takes Longer Than It Should
 
Hi all!

We are running Amanda 3.1.1 under SunOS 5.10 and it seems like the recording phase of our nightly backups is taking longer and longer on certain directories, even though no new additional data has been added to said directories. I have come up with some possible theories as to why this may be happening as well as some possible solutions (using new tapes for the set, cleaning the drive heads, etc). However, I do not think that these solutions would be acceptable to my IT manager. If anyone could shed any light on this, the help would be greatly appreciated.

gmartin 01-16-2012 08:44 AM

Curious as to why your manager wouldn't allow you to clean the drive heads. However, I don't believe those would cause a slow down unless the backup system is reporting errors.

Are you certain the size of the data has not changed?

Are you backing up across a network? Is the network healthy?

Is the data volume healthy? A degraded array due to a bad disk could cause this.

GetLinux 01-16-2012 09:00 AM

Quote:

Originally Posted by robert.wolfe (Post 4572219)
Hi all!

We are running Amanda 3.1.1 under SunOS 5.10 and it seems like the recording phase of our nightly backups is taking longer and longer on certain directories, even though no new additional data has been added to said directories. I have come up with some possible theories as to why this may be happening as well as some possible solutions (using new tapes for the set, cleaning the drive heads, etc). However, I do not think that these solutions would be acceptable to my IT manager. If anyone could shed any light on this, the help would be greatly appreciated.


Tape backups will start all over again if there's a single bit error typically. Where I work we use VTL's virtual tape libraries which are basically large raid arrays that allow the backups to happen and they handle getting the data on tape. You should be able to see this in the logs.

zpaddy 02-02-2012 01:52 PM

Are using amgtar Amanda application to perform Amanda backups?

choogendyk 02-14-2012 07:23 AM

If you haven't solved this problem yet, there are a couple of angles to come at it, depending on how you answer these questions. If it is only a particular Disk List Entry (DLE) that is having troubles, then you should be looking at /var/adm/messages to see if there are scsi errors. Is it consistently the same DLE? Try doing the backup operation of that DLE from the command line and see what it shows. i.e. if you are using ufsdump in Amanda, try a ufsdump from the command line. You can pipe it to some available space through ufsrestore, or pipe it to ufsrestore to /dev/null.

If you are using ZFS, and gnutar or zfs send for backup, then try those individually from the command line.

If, on the other hand, it seems that taping generally is getting slower, then you should try running something to tape from the command line and see what you see. Most tape drives or libraries have either indicator lights or LCD displays that indicate when they think they need a cleaning tape. It is reasonable to use one every once in a while, though the frequency depends on the type of tape drive.

In any case, the general strategies with Amanda are to familiarize yourself with log files and read them when you have troubles, and then to break down the backup and see which step is causing the problem. Amanda uses native utilities on the OS, and you can do the same thing from the command line. That allows you to narrow the problem down and see exactly where it is.


All times are GMT -5. The time now is 06:20 PM.