You can try turning on extended logging inside /kernel/drv/qlc.conf. THis will dump a whole bunch of messages in system log file (/var/adm/ messages). This might tell us if there are any logouts or rscns happening. I assume that the messages from the backup software have a timestamp which can be related back to messages in /var/adm/ messages. We should be able to tell if the system is getting any events from the SAN or is trying to do something.
A little bit more data on system discovery/probe. Mostly There are 2 kind of things that can interrupt a tape backup. A fibre channel login (PLOGI) or some SCSI level command (other than basic inquiry and report lun). SInce I dont know the details of HP backup software, I cant say what is going on at SCSI level for sure but If the HP backup software is already communicating with this device that means no other SCSI driver inside solaris has attached to this device and there cannot be any other SCSI I/O other than what is generated by HP backup software. Regarding FC login, that only happens during initial discovery process and does not happen again unless the device or the fabric has generated an event. Sumit On Oct 28, 2007, at 12:46 PM, Tom De Boeser wrote: >> >> I do not think there is any relation between I/O >> errors and system >> device discovery. The thing to understand is what >> events are >> happening in your SAN which result in I/O errors. Are >> you doing >> anything or causing any events or the errors are >> happening on their >> own. > >> If they are happening on their own then what is >> the frequency of >> the errors (is there any relation to the load e.g. >> the error happens >> 3 min. after the 1st I/O). > > They (i/o err) happen randomly at different times. Sometimes if > we are at work monitoring backups, most times when automated > backups are scheduled over night. I fairly confident the errors > are related system discovery/probe because this has happened in the > past ( fixed by disabling "auto-discovery", and bus monitoring > apps), and because the backups were stable until switching to this > new system. > >> >> Also which driver is generating I/O errors ? (perhaps >> a clipping of / >> var/adm/messages will help). And what is the system >> configuration ? >> i.e. OS level, SAN patch level, 3rd party software, >> HBA, switch etc ? > > There are no errors on the system, the errors are reported from the > backup software. HP tells this error happens when the tape device > is interrupted. The system has been patched including SAN patches, > i'll have to get those a little later. But I downloaded SAN > related patch two weeks ago. > We have Sun's Qlogic HBA's, we were going to try and use Qlogics > drivers, but the qlc driver seems to work, and was suggested over > the qlogic. Besides Data Protector there isn't any other software > installed, umm... maybe someone put the qlogic san software on, > I'll have to look. > > Also, since this is somewhat "normal" behavior the switch(es) and > libraries don't report any problems. > >> >> Sumit > > Thanks, > > Tom de > > > This message posted from opensolaris.org > _______________________________________________ > storage-discuss mailing list > [email protected] > http://mail.opensolaris.org/mailman/listinfo/storage-discuss _______________________________________________ storage-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/storage-discuss
