| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 1206.1 | roger d., retry the errorlog with decevent.. | SUBSYS::VIDIOT::PATENAUDE | Ask your boss for ARRAY's... | Mon Feb 10 1997 13:52 | 6 |
|
$diag/since=08-feb-1997
Anal/err is giving you bogus info "controller memory error"
roger p.
|
| 1206.2 | DIAG? | SUBSYS::DUVAL | | Mon Feb 10 1997 14:59 | 10 |
|
$diag/since=08-feb-1997 ?
$DIAG not a DCL command?
Is decevent something layered on VMS?
thanks,
roger d.
|
| 1206.3 | Yup could be battery | SSDEVO::RMCLEAN | | Mon Feb 10 1997 17:23 | 4 |
| The instance code is cache battery bad. It might be that in the middle of
all of this you lost the cache battery's. Did you have a failover occur also?
The failover would have caused inaccessability. A bad battery would have
caused that too.
|
| 1206.4 | | SUBSYS::VIDIOT::PATENAUDE | Ask your boss for ARRAY's... | Mon Feb 10 1997 18:34 | 7 |
|
Roger,
Yes it's layered but not on LEDS (VAX), I tried it on MSGAXP, but being a lowley
user have no privs to the errorlog ;^) but the command did take.
Roger.
|
| 1206.5 | 2.7 Patch problem now... | SUBSYS::DUVAL | | Tue Feb 11 1997 13:14 | 22 |
| Yes. One controller's battery goes to FAILED every few days. I do a
restart and it comes back to "GOOD". FS pointed me to a patch for
2.7 that may help this situation. I pulled the patch this morning,
applied it to all 4 controllers and now i'm in REAL trouble. After
the patch I did a restart and ever since then I have not been able
to get them to come to life again. I tried reloading the card and
it appears to try, it sequences, then the green light flashes off
and on just as the other does, but then it stays out and never comes
back, nothing on console?
I tried a known good card and same result. Where did i go wrong? No
errors or problems applying the patches. I pulled the patch from the
CSC32:: location. I've applied patches in the past to 2.5 without a
problem?
All my disks are now attached to my 2 controllers that I dare not
restart until I can get this resolved.
Realy need help now,
thanks,
roger d.
|
| 1206.6 | Try this... | SSDEVO::RMCLEAN | | Tue Feb 11 1997 14:34 | 124 |
| I haven't seen this but... Take the patch out of the controllers that are
currently running. Then at least you won't have a problem if they crash.
Next I would strongly suggest you get good batteries. The patch only helps
when you have certain types of batteries that are good but exhibit interesting
properties at certain temperatures. I would log an IPMT case and see if you
can get other insight there....
To ensure you have the right patch I will include a copy below:
Patch Title: Fix for Battery Test failures
HSOF Version: V27J
Patch Number: 1
Date: 24-SEP-1996
Engineer:
I. Patch Description:
The current Periodic Cache Battery Test algorithm does not provide
sufficient test coverage for the controller to properly detect good/bad
Cache Batteries. This patch improves the test coverage provided by the
Periodic Cache Battery Test.
II. Patch in text form:
Extract the following text to PATCH_V27J-1.txt and send to customers entering
the patch by hand.
----- Begin text -----
Title: Fix for Battery Test failures
Version: V27J
Length: 128
Patch Type: 0
Patch Number: 1
Count: 27
Address: 20108310
Value[ 0]: 90383000
Value[ 1]: 20171CC8
Value[ 2]: 58E8198B
Value[ 3]: 90A9E008
Value[ 4]: 8C803000
Value[ 5]: 0027BC86
Value[ 6]: 09EF7EAC
Value[ 7]: 58A01989
Value[ 8]: 92A1E008
Value[ 9]: 8C803000
Value[ 10]: 0013DE43
Value[ 11]: 09EF7E98
Value[ 12]: 90A1E008
Value[ 13]: 305D202C
Value[ 14]: 598E5E03
Value[ 15]: 5947DE10
Value[ 16]: 5936DE10
Value[ 17]: 65320A80
Value[ 18]: 8C803000
Value[ 19]: 20142AE0
Value[ 20]: 0901ED50
Value[ 21]: 90A1E008
Value[ 22]: 8C91D000
Value[ 23]: 375D2010
Value[ 24]: 8C91D000
Value[ 25]: 58840090
Value[ 26]: 305D6038
Count: 0
Verification: 30C80AA1
----- End text -----
III. Patch Installation Script for use with HSDSA-SCRIPT.EXE:
Extract the following script to PATCH_V27J-1.script and execute it using the
HSDSA-SCRIPT program.
----- Begin Script -----
!
!Fix for Battery Test failures
!
run clcp
2
1
y
V27J
128
0
1
27
20108310
90383000
20171CC8
58E8198B
90A9E008
8C803000
0027BC86
09EF7EAC
58A01989
92A1E008
8C803000
0013DE43
09EF7E98
90A1E008
305D202C
598E5E03
5947DE10
5936DE10
65320A80
8C803000
20142AE0
0901ED50
90A1E008
8C91D000
375D2010
8C91D000
58840090
305D6038
0
30C80AA1
3
0
----- End Script -----
|
| 1206.7 | Backed out the v27j-1 patch | SUBSYS::DUVAL | | Wed Feb 12 1997 17:23 | 13 |
| Well I did remove the patch from my one surviving HSJ before It
rebooted. It did turn out to be the patch that caused it, I had
to boot from a v25j card (which ignored the 27j-1 code patch) and
then go in and delete the v27j-1 patch. I did that to 3 controllers
and rebooted them with the origional v27j cards. Then I had to do
a "clear lost_data" for all my units. CSC helped me with all that.
The patch and log included in .6 is exactly what I had done, each
log was clean with no indication of a problem?
thanks,
roger d.
|
| 1206.8 | Status / | GIDDAY::HOBBS | Andy Hobbs. Sydney CSC. -730 5964 | Tue Feb 18 1997 00:48 | 10 |
|
Made any progress in understanding this situation, Roger ?
I've not seen it myself and I've pushed this patch onto a few
controllers (Australia and UK) without any hitches before, during
or after.
How did you perform the post-patch reboots ?
A/.
|
| 1206.9 | bad v27j.img file for patching anyway... | SUBSYS::DUVAL | | Fri Mar 07 1997 11:51 | 14 |
| Hi,
Yes. This one has been understood. I had an unofficial v27j-0 release
to start with. I received 4 Hsj controllers with no cards last year,
and never could convince them to send me any, so I had to beg. We had
a blaster here in SHR, so all I needed was the v27j.img file. The one
I got was a bit off for applying a patch (although the patch utility
did not flag a problem). The patch was corrupting the OS and so a
"restart" would just hang.
I re-blasted all my cards and the patch works fine now...
thanks,
roger d.
|