The last log backup for volume 5 (indexserver on hana2:30003) failed.

[Former Member]

Experts

I am getting this alert in our HANA Production DB. The Log backup for volume 5 failed. This was happening since many days and could not find the answer from SMP.

The last log backup for volume 5 (indexserver on hana2:30003) failed.

There is no error reported from Backup Console.  We have four server nodes and volume 5 is our Index Server on Host2.

your thoughts would be appreciated

Mahesh Shetty

[SK-EA]

Mahesh,

What is the appliance make and model? If its IBM then look into quotas. Even though you see free space, the quotas may run out of space.

Hope this help.

Regards,

Srinivas K

[Former Member]

How are the log backups configured?

Do you have a native HANA log backups or BACKINT setup?

Is this an intermittent failure or recurring very often?

For native backup failures check the backup.log under the HANA STUDIO-->DIAGNOSIS FILES

For BACKINT setup we have an additional backint.log under Diagnosis files

These are the right points to start the investigation

[Former Member]

We have Space left on Device and also regular backups are going fine. The Log backups for other volumes are fine just this volume it complains. We are using File backups in our Backup configuration.

I dont see any errors in the backup.log file since Mar 15th 2014. That was the last time we had a successfull Log backup.

2014-03-15T00:13:55-05:00  P014357      144c428ff41 INFO    LOGBCKUP state of service: indexserver, hana2:30003, volume: 5, BackupLogSuccess

2014-03-15T00:13:56-05:00  P014357      144c4292a1e INFO    LOGBCKUP state of service: indexserver, hana2:30003, volume: 5, BackupLogStarted

2014-03-15T00:13:56-05:00  P014357      144c4292a1e INFO    LOGBCKUP to file: /usr/sap/HDP/HDB00/backup/log/log_backup_5_0_8951293184_8959322624.1394860435998

2014-03-15T00:13:56-05:00  P014357      144c4292a1e INFO    LOGBCKUP start of progress monitoring, volumes: 1, bytes: 513884160

2014-03-15T00:13:56-05:00  P014357      144c4292a1e INFO    LOGBCKUP progress of service: indexserver, hana2:30003, volume: 5, 0% 0/513884160

2014-03-15T00:13:57-05:00  P014357      144c4292a1e INFO    LOGBCKUP progress of service: indexserver, hana2:30003, volume: 5, 26% 134217728/513884160

2014-03-15T00:13:59-05:00  P014357      144c4292a1e INFO    LOGBCKUP progress of service: indexserver, hana2:30003, volume: 5, 52% 268435456/513884160

2014-03-15T00:14:00-05:00  P014357      144c4292a1e INFO    LOGBCKUP progress of service: indexserver, hana2:30003, volume: 5, 78% 402653184/513884160

2014-03-15T00:14:01-05:00  P014357      144c4292a1e INFO    LOGBCKUP progress of service: indexserver, hana2:30003, volume: 5, 100% 513884160/513884160

Dont see any logbackups for Indexserver on volume for hana2 host. This is strange and am I missing something here. The Log file is very big and hence cannot attach it to the Post.

Mahesh Shetty

[Former Member]

Strange to see the backup.log has no errors since more than 1 year although the log backup fails

Where do you find the error you posted in the first message?

When did you open the incident with SAP Support?

I would like you to check this:

We have a setting for log_mode under the global.ini--> [persistence]

I notice we have an option to set a SYSTEM level log_mode setting from HANA Studio

Can you check if your distributed Node 2 has a log_mode set to OVERWRITE which may be causing this issue? That may explain why the log backups do not occur for the Node 2 Indexserver alone

[Former Member]

Thanks Sunil

We are using log_mode=overwrite so far as we have not setup any backint agent and also we never went live with any HANA projects so far. We have one project scheduled to go live this month end and we are working on Log backups using backint for point of recovery.

Yes, the setting mentioned as been set to overwrite for parameter log_mode under global.ini --> persistence.

Have a quick question. Why does it fail only for Indexserver on this particular node when we have other nodes in the system. I appreciate your thoughts on this.

Mahesh Shetty

[Former Member]

Interesting so lets be clear on the below points:

log_mode=OVERWRITE can be set for non prod HANA systems to ensure we do not have any log backups taken on them. This setting has no relation to the use of BACKINT or NATIVE HANA backups, please refer Administration guide for this recommendation

Now can you share a screen shot of the log_mode setting on your side?

I am interested to know if log_mode is set to NORMAL on SYSTEM level and OVERWRITE on HOST level

Ater capturing this, please change the log_mode setting to NORMAL under SYSTEM in HANA Studio

This will ensure all services logs are backed up as expected

log_mode change can be done online and does not need a restart, however please take a Data backup as soon as you switch the log_mode to NORMAL

Let me know if this helps resolve the issue

[Former Member]

Could you please check the if the bacckup DIR is accessible.Also - trace will help to better understand the issue.

Thanks

Dev

[Former Member]

Paste the Indexserver Log also check if you have enough space

[former_member182967]

Hello Mahesh,

Since you don't provide the trace files, check if the following notes help or not:

1905173 - Backup of a particular indexserver fails in HANA distributed environment

2058258 - HANA database backup fails with "allocation failed" error

Regards,

Ning Tong