Networker NMDA asdf_output_section1() failed

Hello,

I have found this error on one database backup.
There are several databases running on the same db server with same oracle home, same rman backup script (just different db name). Networker version is 9.2.1.5 on both server and backupserver. Backup is on the datadomain devices. 

There is only one failing database with following error:

RMAN-03009: failure of backup command on t1 channel at 01/08/2020 12:22:09
ORA-19502: write error on file "tiulf9og_2_1", block number 16097 (block size=8192)
ORA-27030: skgfwrt: sbtwrite2 returned error
ORA-19511: non RMAN, but media manager or vendor specific failure, error text:
asdf_output_section1() failed
xdr=0x1143dede0: bp=0x1144acaa8: send_len=262144: type=12800: fhand=0x1144100e8: wrapper=0x0: directp=0xffffffff7694f000 (1:4:32)

To isolate the issue we have run separate backups of datafiles and it seems this only happens on the biggest one (1.3G). It seems that once the backup hits 512-13MB it just hangs for a few minutes then fail. We split the backup to smaller chunks (500mb) and this time the first backup piece finished and the second one failed. Suspected some problem on filesystem but filesystem check did not return any errors. 
On the other databases on this host we also have more than 1GB datafiles so I presume it is not some kind of blocksize issue or timeout. Also this happens only with scheduled backups, manual backup using the same rman script is successful.
Archive log backus are ok on this database, only L0 and L1 are failing
Checked also datadomain devices and did not find any problem (ran backup to tape as well to check but same error).

Apart from basic trobleshooting I tried to restart networker, reinstall networker and nmda (cleared whole nsr on the client), cleared NSR peer info also checked the keepalive params, no idea where to look next.