iSCSI on 8.04 lTS

[Kent Nasveschuk]

Hello,
I recently installed Ubuntu 8.04 LTS on a Dell 2950 dual proc quad core. I am using iSCSI/multipath over an Intel dual port nic to connect to a Compellent SAN. Yesterday I had both paths (multipath) crash. Here is the output from kern.log:

Dec  9 22:00:47 mail kernel: [484935.820115] sd 8:0:0:0: [sdd] Result: hostbyte=DID_ABORT driverbyte=DRIVER_OK,SUGGEST_OK
Dec  9 22:00:47 mail kernel: [484935.820123] end_request: I/O error, dev sdd, sector 213888
Dec  9 22:00:47 mail kernel: [484935.820126] device-mapper: multipath: Failing path 8:48.
Dec  9 22:00:50 mail kernel: [484938.529252] device-mapper: multipath: Failing path 8:32.
Dec  9 22:01:17 mail kernel: [484965.803624] sd 7:0:0:0: [sdc] Result: hostbyte=DID_ABORT driverbyte=DRIVER_OK,SUGGEST_OK
Dec  9 22:01:17 mail kernel: [484965.803633] end_request: I/O error, dev sdc, sector 12884901760
Dec  9 22:01:17 mail kernel: [484965.803641] Buffer I/O error on device dm-0, logical block 1610612720
Dec  9 22:01:17 mail kernel: [484965.803698] sd 7:0:0:0: [sdc] Result: hostbyte=DID_ABORT driverbyte=DRIVER_OK,SUGGEST_OK
Dec  9 22:01:17 mail kernel: [484965.803703] Buffer I/O error on device dm-0, logical block 1610612720
Dec  9 22:01:17 mail kernel: [484965.803705] end_request: I/O error, dev sdc, sector 213888
Dec  9 22:01:17 mail kernel: [484965.803761] Buffer I/O error on device dm-0, logical block 1610612734
Dec  9 22:01:17 mail kernel: [484965.803773] Buffer I/O error on device dm-0, logical block 26742
Dec  9 22:01:17 mail kernel: [484965.803777] lost page write due to I/O error on dm-0
Dec  9 22:01:17 mail kernel: [484965.803780] Aborting journal on device dm-0.
Dec  9 22:01:17 mail kernel: [484965.803785] EXT3-fs error (device dm-0) in ext3_dirty_inode: IO failure
Dec  9 22:01:17 mail kernel: [484965.803791] Buffer I/O error on device dm-0, logical block 1289
Dec  9 22:01:17 mail kernel: [484965.803794] lost page write due to I/O error on dm-0
Dec  9 22:01:17 mail kernel: [484965.803903] Buffer I/O error on device dm-0, logical block 1610612734
Dec  9 22:01:17 mail kernel: [484965.803963] Buffer I/O error on device dm-0, logical block 1610612735
Dec  9 22:01:17 mail kernel: [484965.803997] Buffer I/O error on device dm-0, logical block 1610612735
Dec  9 22:01:17 mail kernel: [484965.804038] Buffer I/O error on device dm-0, logical block 1610612735
Dec  9 22:01:17 mail kernel: [484965.804074] Buffer I/O error on device dm-0, logical block 1610612735
Dec  9 22:01:17 mail kernel: [484965.804088] ext3_abort called.
Dec  9 22:01:17 mail kernel: [484965.804091] EXT3-fs error (device dm-0): ext3_journal_start_sb: Detected aborted journal
Dec  9 22:01:17 mail kernel: [484965.804094] Remounting filesystem read-only
Dec  9 22:01:47 mail kernel: [484995.924803] printk: 22 messages suppressed.
Dec  9 22:01:47 mail kernel: [484995.924809] Buffer I/O error on device dm-0, logical block 578388174
Dec  9 22:01:47 mail kernel: [484995.924841] lost page write due to I/O error on dm-0
Dec  9 22:01:47 mail kernel: [484995.924848] Buffer I/O error on device dm-0, logical block 578392064
Dec  9 22:01:47 mail kernel: [484995.924877] lost page write due to I/O error on dm-0
Dec  9 22:01:47 mail kernel: [484995.924883] Buffer I/O error on device dm-0, logical block 843579394
Dec  9 22:01:47 mail kernel: [484995.924913] lost page write due to I/O error on dm-0
Dec  9 22:01:47 mail kernel: [484995.924917] Buffer I/O error on device dm-0, logical block 843585538

You can see the file system was marked "ready only". This has only been running about a week.

I've updated 8.04 with latest updates. Would like some help on resolving the issue and/or making the iSCSI connection more reliable.

I have Redhat and CentOS boxes using the SAN via iSCSI reliably. 




Kent L. Nasveschuk
Systems Administrator

----------------------------
Marine Biological Laboratory
7 MBL St.
Woods Hole, MA 02543