BUG REPORT: Large NDMP backups to filers fail with status code 23 (socket read failed) reported.

Problem

BUG REPORT: Large NDMP backups to filers fail with status code 23 (socket read failed) reported.

Error Message

write_data: ndmp_xm_get_kbytes failed

Solution

Bug: 1195987

Detail/Symptoms:
Very large NDMP backups to filer fail with status 23 (socket read failed) reported.

Log Files:
In the bptm log of the media server, messages similar to the following are found:
19:01:10.180 [4792.5372] <2> NdmpSession: [514] Sending 15 (GET_KBYTES) ""
19:06:07.258 [4792.5372] <2> NdmpSession: [514] Reply 15 "433281744", error = 0
19:06:37.258 [4792.5372] <2> NdmpSession: [515] Sending 15 (GET_KBYTES) ""
19:11:37.790 [4792.5372] <2> NdmpSession: [515] Reply 0 "", error = 18
19:11:37.790 [4792.5372] <16> write_data: ndmp_xm_get_kbytes failed

In reality, the ndmpagent process is still working (and its logs will confirm this as messages will continue to be logged past the reported error), but bptm doesn't "get the message" - and after the default timeout of five minutes (300 seconds), a failure is reported.

Resolution:
The NDMP_CONNECT_TIMEOUT touch file can be used to increase the value of this timeout.  It will need to be located in:
Windows: <install_path>\NetBackup\ndmp_connect_timeout
UNIX: /usr/openv/netbackup/ndmp_connect_timeout

This file should contain the value in seconds of the desired timeout.  For instance, a value of 480 would correspond to an 8 minute timeout.

This functionality was introduced in the following release:
  • NetBackup 6.0 Maintenance Pack 7 (MP7)

To see if this change has taken effect, check the ndmp logs (OID 151) for the following message:
ndmp_net_connect: Timeout Value: 480000

The timeout will be specified in microseconds (that is, the value of NDMP_CONNECT_TIMEOUT multiplied by 1000).

Note: Starting in NetBackup 6.5, the value of CLIENT_READ_TIMEOUT is used to set the NDMP connect timeout.  The above mentioned touch file is not used.  If this condition is seen in a NetBackup 6.5.x environment, please adjust the client connect timeout value to resolve this issue.

Terms of use for this information are found in Legal Notices.

Search

Survey

Did this article answer your question or resolve your issue?

No
Yes

Did this article save you the trouble of contacting technical support?

No
Yes

How can we make this article more helpful?

Email Address (Optional)