need help to under stand infomation in messages logfile
 

have a system that is linux RHE 4.0 with a rdms postgresql ver 7.4.8 running on it. Not by choice the system is also setuyp with RAID 5. two of the postgress apps vacuumdb and pg_dump havine been failing due to an I/O error of some sort. A posting of the postgresql user forum got me the anser thatwer that its some type of hardware problem and to check the system logfiles. The following is an excerpt from the messages logfile. Can someone please tell me what it means. Is there a hardware problem or not and if so does it tell or hint at what type of hardware problem it is. The elec. technician says the postgres answer is hot air that they always say hardware.

Jul 8 09:40:12 ax-orn postgres[31684]: [522-1] INFO: "peoosep": 350762 pages, 3000 rows sampled, 16836576 estimated total rows
Jul 8 09:40:13 ax-orn postgres[31684]: [523-1] INFO: vacuuming "public.unkstnvalue"
Jul 8 09:42:10 ax-orn kernel: SCSI error : <1 0 0 0> return code = 0x70000
Jul 8 09:42:10 ax-orn kernel: end_request: I/O error, dev sda, sector 31255429
Jul 8 09:42:10 ax-orn kernel: SCSI error : <1 0 0 0> return code = 0x70000
Jul 8 09:42:10 ax-orn kernel: end_request: I/O error, dev sda, sector 31255437
Jul 8 09:42:10 ax-orn kernel: SCSI error : <1 0 0 0> return code = 0x70000
Jul 8 09:42:10 ax-orn kernel: end_request: I/O error, dev sda, sector 31255437
Jul 8 09:42:10 ax-orn postgres[31684]: [524-1] ERROR: could not read block 83 of relation "unkstn_time_ind": Success
Jul 8 09:42:21 ax-orn crond(pam_unix)[31652]: session closed for user oper
Jul 8 09:44:15 ax-orn postgres[32572]: [57-1] LOG: recycled transaction log file "00000025000000DF"

What all does this mean?
Any information is greatly appreciated as I and the other person that's been trying to figure this our are stumped.

Julie

Hi,

and welcome to LQ!

Yes, there's definitely an issue. I don't know anything
else about your set-up, but assuming that this is a
hardware RAID5, and all the drives in the array are
mapped to one logical drive this is really bad, because
it indicates that the problem spreads over all drives.


Cheers,
Tink