IBM iSeries Series Hardware Problem Analysis And Isolation page 93

5. Switch the two bus cables on the second I/O tower's failed port so that the cable that was connected
to port 1 will now be in port 0, and the cable that was connected to port 0 will now be in port 1.
Attention: You must fully connect the cable and tighten the connector's screws within 30 seconds of
when the cable makes contact with the port. Otherwise, the link will fail and you must disconnect and
reconnect again. Also, if the connector screws are not tightened, errors will occur on the link and it
will fail.
6. Refresh the port status for the first failing resource by doing the "Refresh the port status" on page 82
subtask in this document, then continue with the next step.
7. If the port on the I/O tower on which you did NOT switch cables is now working, use the symbolic
FRU "SIIOADP" on page 450 to exchange the HSL I/O bridge card in the I/O tower on which you DID
switch the cables. Go to Verify the repair.
8. Switch the cables back to their original positions.
Attention: You must fully connect the cable and tighten the connector's screws within 30 seconds of
when the cable makes contact with the port. Otherwise, the link will fail and you must disconnect and
reconnect again. Also, if the connector screws are not tightened, errors will occur on the link and it
will fail.
9. Exchange the cable between the two ports on the failed link. Go to Verify the repair.
10. If you are working on a system Model 800, 810, 825, 870, or 890, then use symbolic FRU
"BUSCBL1, BUSCBL2, BUSCBL3, or BUSCBLx" on page 292. For all other systems, use the
symbolic FRU "SIRGCBL" on page 453 to replace other cable related items that may be on either
frame, such as interposer cards and cable wraps. If the error was cleared up by exchanging these
items, go to Verify the repair. If the error still has not been cleared up, call your next level of support
for further instructions.
Cannot power on frame:
v If one or more of the frames in the loop cannot become powered on, work the errors related to
powering on the frames before continuing with this bus error.
v If a frame still cannot become powered on, re-configure the HSL loop without the I/O towers and system
units that cannot be powered on, allowing the loop to complete.
v If the error was cleared up when the frames were powered back on or removed from the loop, go to
Verify the repair.
v If the error persists when the frames were powered back on, or removed from the loop, continue from
where you left off in the "Main task" on page 79.
Manually detecting the failed link:
1. Get the loop number from the SRC if you do not already have it. The loop number is a hexadecimal
number in word 7 of the SRC. If you are working with the SRC in the PAL, then the loop number is
the 4 leftmost characters of the DSA in word 7 (BBBB). Use the "DSA breakdown" on page 63 (in the
"Bus, high-speed link (HSL) PIPs" on page 62 overview section) to convert the hexadecimal loop
number to decimal format before continuing with this procedure.
If you are working from the Service Action Log (SAL), the loop number should be displayed in the
FRU description area in decimal format.
2. Sign on to SST or DST (if you have not already done so). Select Start a service tool —> Hardware
service manager —> Logical hardware resources —> High-speed link (HSL) resources.
3. Select Resources associated with loop for the HSL loop with the failed link. The HSL bridges will
be displayed under the loop.
4. Select Display detail for the loop with the failed link.
5. Select Follow leading port until a port with a "failed" status is found. Note the resource name at the
leading port with a failed status, and the type-model and serial number for the resource with the failed
status.
6. Select Follow leading port one more time and note all the information for the resource name with a
failed trailing port.
Analyze hardware problems
81