Error Message

vmcore-dmesg.txt shows as below:

<3> 47212th index is not actually written, pending to be written.
<3> 47221th index is not actually written, pending to be written.
<3>No empty position for the Sectors in the array

<1>BUG: unable to handle kernel NULL pointer dereference at (null)

<4>Call Trace:
<4> [] bio_endio+0x1d/0x40
<4> [] osBioEndio+0xe/0x10 [symbdsnap]
<4> [] snapAccessMakeRequest+0x28c/0x2bb [symbdsnap]

While a snapshot is available, the snapshot driver symbdsnap will save the original data when a write operation occurs. This is caused by insufficient free space to save the original data.

Perform any of the following

1. Set the saving space (a.k.a. cache) to a volume which has enough free space.

Edit the following two entries, both in /etc/VRTSralus/ralus.cfg.

• Software\Veritas or Symantec\Backup Exec For Windows\Backup Exec\Engine\RALUS\AOFOL\CacheFileMountPoint=

The alternate cache file (saving space) location. This location should be the mount point. By default, the snapshot cache file is created on the mount point of the volume that you plan to back up.

• Software\Veritas or Symantec\Backup Exec For Windows\Backup Exec\Engine\RALUS\AOFOL\CacheFileSize=

The initial cache file size. The size should be at least 10% of the volume size that you plan to back up. The unit of this value is 4 KB blocks. For exapmple 40000000 means 152 GB.

2. Disable AOFO for the backup jobs.

3. Run the backup job while the system is not busy (no disk writing)

4. Separate the backup job to 2,3 jobs or more.

This would shorten the backup time of a job, then reduce the amount of data to be written while a snapshot is available.

 

RedHat Enterprise Linux server (RHEL) kernel may crash during AOFO enabled backups.

Error Message

vmcore-dmesg.txt shows as below: <3> 47212th index is not actually written, pending to be written.
<3> 47221th index is not actually written, pending to be written.
<3>No empty position for the Sectors in the array <1>BUG: unable to handle kernel NULL pointer dereference at (null) <4>Call Trace:
<4> [<ffffffff811b954d>] bio_endio+0x1d/0x40
<4> [<ffffffffa05b542e>] osBioEndio+0xe/0x10 [symbdsnap]
<4> [<ffffffffa05a4c55>] snapAccessMakeRequest+0x28c/0x2bb [symbdsnap]

Cause

While a snapshot is available, the snapshot driver symbdsnap will save the original data when a write operation occurs. This is caused by insufficient free space to save the original data.

Solution

Perform any of the following

1. Set the saving space (a.k.a. cache) to a volume which has enough free space.

Edit the following two entries, both in /etc/VRTSralus/ralus.cfg.

• Software\Veritas or Symantec\Backup Exec For Windows\Backup Exec\Engine\RALUS\AOFOL\CacheFileMountPoint=

The alternate cache file (saving space) location. This location should be the mount point. By default, the snapshot cache file is created on the mount point of the volume that you plan to back up.

• Software\Veritas or Symantec\Backup Exec For Windows\Backup Exec\Engine\RALUS\AOFOL\CacheFileSize=

The initial cache file size. The size should be at least 10% of the volume size that you plan to back up. The unit of this value is 4 KB blocks. For exapmple 40000000 means 152 GB.

2. Disable AOFO for the backup jobs.

3. Run the backup job while the system is not busy (no disk writing)

4. Separate the backup job to 2,3 jobs or more.

This would shorten the backup time of a job, then reduce the amount of data to be written while a snapshot is available.

ETrack: 3955714