| Title: | AdvFS Support/Info/Questions Notefile |
| Notice: | note 187 is Freq Asked Questions;note 7 is support policy |
| Moderator: | DECWET::DADDAMIO |
| Created: | Wed Jun 02 1993 |
| Last Modified: | Fri Jun 06 1997 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 1077 |
| Total number of notes: | 4417 |
Hi!
Is it possible that in a domain which ist built up with more
than one physical partition (using the addvol-command), a write
operation may complete, even if one of the disk (partitions) is
offline?
for example:
i have a domain with two physical partitons named "test_domain"
# ls -l /etc/fdmns/test_domain
lrwxr-xr-x 1 root system 10 Jan 29 21:31 rz8c ->
/dev/rz8c
lrwxr-xr-x 1 root system 10 Jan 29 21:31 rz9c ->
/dev/rz9c
lets imagine my operating system writes to a file which is
perhaps only located on rz8c. Now the disk rz9c fails for some reason
(hardware fault). Will the actual write operation of my file complete
since it only accesses rz8c or will the domain panic immediately?
Helmut
| T.R | Title | User | Personal Name | Date | Lines |
|---|---|---|---|---|---|
| 998.1 | Should succeed... | NETRIX::"[email protected]" | Arti Kumar | Wed Feb 05 1997 15:36 | 11 |
The write to a file that is located entirely on one disk should succeed. If the other (failing) disk is offline, an access to files on that disk will return an error. If the failing disk is still online then the next metadata access to it will cause a domain_panic, and the next user access to it will return an error. If you notice that one of the disks in a multi-volume domain is returning errors or is failing you should backup the domain, and then remove the offending disk using rmvol. [Posted by WWW Notes gateway] | |||||
| 998.2 | thanks | VNABRW::PUTZENLECHNE | Thu Feb 06 1997 04:54 | 15 | |
Thank You!
The reason i asked these questions was that we lost a database
because one disk within a domain went offline and after
bringing it back online the database (oracle) was inconsistent.
We try to find out the reason for this behaviour together with
oracle, because normally the database should be able to recover from
such hardware-failures.
regards,
Helmut
| |||||