https://bugzilla.novell.com/show_bug.cgi?id=366492 Summary: file system corruption and scsi errors with open-iscsi and Infortrend iSCSI arrays Product: SUSE Linux 10.1 Version: Final Platform: x86-64 OS/Version: SLES 10 Status: NEW Severity: Critical Priority: P5 - None Component: Network AssignedTo: bnc-team-screening@forge.provo.novell.com ReportedBy: s.pigneguy@niwa.co.nz QAContact: qa@suse.de Found By: Customer Starting with a fresh installation of SLES 10, and upgrading to SP1 I can reliably get open-iscsi to corrupt a filesystem and return scsi errors with an Infortrend iSCSI array. The problem goes away when I use the latest stable release of open-iscsi from the open-iscsi website. Previous to using the latest version of open-iscsi I had replaced all our network switches to quality cisco equipment, reinstalled & reloaded servers, reformatted the iSCSI arrays (we have two identical arrays). The problem version of open-iscsi is open-iscsi-2.0.707-0.24 on both i386 & x86_64 The version of open-iscsi that works is open-iscsi-2.0-865.15 The scsi errors I see are like : (summary of dmesg) sd 14:0:0:1: SCSI error: return code = 0x00020000 end_request: I/O error, dev sdj, sector 3379268680 (summary of /var/log/messages) Feb 26 12:59:19 thor kernel: Buffer I/O error on device dm-9, logical block 67363007 Feb 26 12:59:19 thor kernel: lost page write due to I/O error on dm-9 Feb 27 12:46:43 thor kernel: end_request: I/O error, dev sdj, sector 2041855152 Feb 27 12:46:43 thor kernel: I/O error in filesystem ("dm-9") meta-data dev dm-9 block 0x79b43b30 ("xfs_trans_read_buf") error 5 buf count 8192 Feb 27 12:46:44 thor iscsid: connection1:0 is operational after recovery (1 attempts) The Infortrend iSCSI Array model is : A16E-G2130-4 The network switch model is : Cisco 2960G The server models are : HP DL380, HP DL360 Thanks, Stuart. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.