EBU-7042 Sbtwrite: I/O error

[cbrae]

We are using Legato 5.5.2 Build 165 and obackup (Oracle version 7.3.4.3) and on a fairly frequently getting Sbtwrite errors. These are 40 gig or so DBs ... our smaller DBs are doing just fine .. any ideas/tuning or ?? that may help relieve this ..<br><br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;thanks ... <br><br>

 

[jisler2]

We are also receiving this same error on our site.<br>i.e. EBU-7042: Sbtwrite: I/O error.<br>We have the NetWorker 5.1.2 Extended(Power Edition)/13 using 40Gb DLT.<br>Database size is approx 100Mb only.<br>We are frequently receiving this error and consequently backup fails.<br><br>Any help will be appreciated.<br>Thanks.

 

[cbrae]

The following seems to help -- our incidence seems to have gone down but not ended<br><br>set these environmental variables the big one seems to be BRD_TIMEOUT<br><br>BRD_ERR_TIME=1<br>BRD_RETRIES=1<br>BRD_SAMP_TIME=1<br>BRD_TOT_TIME=300<br>export them <br><br>on nwadmin under media / jukebox - select details<br><br>then set eject sleep to 5<br>cleaning delay 60<br>unload sleep 5<br>load sleep 5<br>deposit timeout 300<br><br>if brd process dies - because it times out - it times out if it doesn't get a request in NN <br>seconds it will go away and you will get an I/O error we upped it to 300 seconds so that on a busy system we should get a request from obackup in 5 minutes - this seemed to minimize our problems .... plus oracle has a new release of obackup for<br>7.3.4 so you might want to go to metalink and look at installing it <br><br>we also seemed to have some problems on changing tapes in slots during the backup<br>timing out on the new tape deposit so we upped it to 300 also