bentele@mtu-friedrichshafen.com writes:
Replacing motherboard and FC-Adapter-Card at last solved the problem, after we replaced all cables, all shelves of this loop, and about 3 disks.
One of my main complaints about the current fcal implementation for the filers is that there is no visual indicator of loop status. I know NetApp is working toward addressing this, but it would have been nice to have from the start.
I wish fcal were more like token ring or fddi and that the loop were "self healing". When I had an LRC fail on an fcal shelf, there was no way of finding which LRC (used process of elimination) was the culprit, as the entire loop was offline, not just the failed shelf and those after it.
Alex
We purchased three F740s which have now been in service for 3 months. We bough an FCAL shelf for one filer but we transferred old SCSI shelves onto the other two. We assumed that we would just get the motherboard FCAL connection, but when they arrived each filer had another two FCAL interfaces in adapter slots 6 and 7. The connectors are completely different from the motherboard FCAL interface.
Unfortunately Netapp did not supply any terminators for these cards, so whenever we boot the filers we get moans like:
Loop on adapter 6 is open. If adapter is in use, check cabling and seating of LRC cards in FC shelves. isp2100_init: Initialization failed on FC adapter in slot 6.
We get this for slots 6 and 7 on the SCSI-only filers, and for the one unused interface on the FCAL filer.
I have moaned a number of times to our UK reseller but they say that Netapp don't supply terminators and we could always pull the cards out. I guess this is true but it's quite useful to have them there as a resilience feature. There are also rumours that these cards are somehow "better" than the motherboard interface.
Can anyone shed any light on this?
Dave Atkin ------------------------------------------------------ Dave Atkin, Head of Technical Services Computing Service, University of York, YORK YO10 5DD Phone: +44-1904-433804 (ddi) Fax: +44-1904-433740 Email: D.Atkin@york.ac.uk ------------------------------------------------------
There is a bug in the firmware of the on-board FCAL controller. NetApp was taking proactive stance on installing FCAL cards into filers and placing a terminator on the on-board controller. As far as I know, there are now terminators for the FCAL card. You might try inserting the cable and then a 9-pin gender changer and then a FCAL terminator. I have never tried this, just live with the warning messages myself, but it sounds like it should work.
-gdg
Dave Atkin wrote:
We purchased three F740s which have now been in service for 3 months. We bough an FCAL shelf for one filer but we transferred old SCSI shelves onto the other two. We assumed that we would just get the motherboard FCAL connection, but when they arrived each filer had another two FCAL interfaces in adapter slots 6 and 7. The connectors are completely different from the motherboard FCAL interface.
Unfortunately Netapp did not supply any terminators for these cards, so whenever we boot the filers we get moans like:
Loop on adapter 6 is open. If adapter is in use, check cabling and seating of LRC cards in FC shelves. isp2100_init: Initialization failed on FC adapter in slot 6.
We get this for slots 6 and 7 on the SCSI-only filers, and for the one unused interface on the FCAL filer.
I have moaned a number of times to our UK reseller but they say that Netapp don't supply terminators and we could always pull the cards out. I guess this is true but it's quite useful to have them there as a resilience feature. There are also rumours that these cards are somehow "better" than the motherboard interface.
Can anyone shed any light on this?
Dave Atkin
Dave Atkin, Head of Technical Services Computing Service, University of York, YORK YO10 5DD Phone: +44-1904-433804 (ddi) Fax: +44-1904-433740 Email: D.Atkin@york.ac.uk
We're running the filer-mrtg graphing program to monitor our filers. One of the measurements graphed is 'Network Kbytes Sent and Received'. For a 100Mb link the theoretical maximum should be about 12.5K KBytes. We're running at 6K KBytes. While this is within spec we're upgrading it to a 1Gb link.
However, filer-mrtg also has the measurement 'Network Interface e10a (type=ethernet-csmacd) - Sent and Received'. For this we are seeing 1280K KBytes peaks. What exactly is this measuring, and what is the maximum value?
Dave
Dave Heiland wrote:
However, filer-mrtg also has the measurement 'Network Interface e10a (type=ethernet-csmacd) - Sent and Received'. For this we are seeing 1280K KBytes peaks. What exactly is this measuring, and what is the maximum value?
I've run into this. The distribution on the NOW website has some default values that are a bit low. In the "conf" directory of your filer-mrtg installation edit the filername.cfg file and change the MaxBytes parameter to something 10 or 100 times larger than what it is. For example, on this filer named "nasty" with a GB ethernet I put it way up there:
Target[nasty.net.interface1]: 1:notreallymysnmpcommunity@nasty MaxBytes[nasty.net.interface1]: 125000000
Graham
-
alexei＠mindspring.net -
Dave Atkin -
Dave Heiland -
G D Geen -
Graham C. Knight