|
Home > Archive > Voice over IP Cisco > October 2007 > 5.1 active disk filling?
You are viewing an archived Text-only version of the thread.
To view this thread in it's original format and/or if you want to reply to
this thread please [click here]
| Author |
5.1 active disk filling?
|
|
| Andre Beck 2007-10-08, 7:11 am |
| Hi,
I didn't have significant problems with 5.1 so far (other than those with
localization that got fixed) and was running 5.1.1.3000-5 until some
weeks ago. Then I upgraded to 5.1.2.3000-2 without problems and the box
ran flawless until yesterday, when CCM suddenly dropped all H.323 gateway
connections. Restarting the CCM service helped.
This morning, the box stopped answering HTTP connections (as tracked
by our Nagios monitoring) and had to be rebooted, after this it again
resumed normal operation.
What I noticed in the OS Admin Web interface is that the "Disk/active"
filesystems are rather full (my 5.1.1 is at 93%, the 5.1.2 was at 95%)
and they are *constantly* *growing*.
I switched versions back to 5.1.1 to verify this, it's the same there:
Total Free Used
Disk/active 12317944K 928084K 10764132K (93%)
[...]
Disk/active 12317944K 928068K 10764148K (93%)
[...]
Disk/active 12317944K 928064K 10764152K (93%)
without the number ever getting smaller again.
What would write to the software partions instead of to the common or
log partitions? Is this normal and maybe garbage collected every some
hours/days or might this be a reason for an installation suddenly failing
after weeks of runtime?
I'm probably going to upgrade to 5.1.3, which will let me lose my 5.1.1
fallback installation. Hopefully this will be stable again.
BTW, the absence of an "expert" login to actually search for the reason
of the growing diskspace is an ENTIRELY SILLY IDEA. I'm feeling like a
lobotomized ex-hacker at this "admin" shell...
Thanks,
Andre.
--
.sig making fun of Santa Claus Operation currently unavailable
-> Andre Beck +++ ABP-RIPE +++ IBH Prof. Dr. Horn GmbH, Dresden <-
| |
| Ryan Ratliff 2007-10-08, 1:12 pm |
| Until you start getting RTMT alerts about the active partition being
full don't worry about it. 93% utilization is pretty normal for a CM
5 box.
There were however some bugs that caused the active partition to fill
up. CSCsk18010 is one of them, but only affects IBM servers and is
fixed in 5.1(3).
I'd encourage you to open a case with TAC to troubleshoot Tomcat (the
web server) and your h.323 issue. They definitely sound like
problems but disk utilization is not the cause.
-Ryan
On Oct 8, 2007, at 4:47 AM, Andre Beck wrote:
Hi,
I didn't have significant problems with 5.1 so far (other than those
with
localization that got fixed) and was running 5.1.1.3000-5 until some
weeks ago. Then I upgraded to 5.1.2.3000-2 without problems and the box
ran flawless until yesterday, when CCM suddenly dropped all H.323
gateway
connections. Restarting the CCM service helped.
This morning, the box stopped answering HTTP connections (as tracked
by our Nagios monitoring) and had to be rebooted, after this it again
resumed normal operation.
What I noticed in the OS Admin Web interface is that the "Disk/active"
filesystems are rather full (my 5.1.1 is at 93%, the 5.1.2 was at 95%)
and they are *constantly* *growing*.
I switched versions back to 5.1.1 to verify this, it's the same there:
Total Free Used
Disk/active 12317944K 928084K 10764132K (93%)
[...]
Disk/active 12317944K 928068K 10764148K (93%)
[...]
Disk/active 12317944K 928064K 10764152K (93%)
without the number ever getting smaller again.
What would write to the software partions instead of to the common or
log partitions? Is this normal and maybe garbage collected every some
hours/days or might this be a reason for an installation suddenly
failing
after weeks of runtime?
I'm probably going to upgrade to 5.1.3, which will let me lose my 5.1.1
fallback installation. Hopefully this will be stable again.
BTW, the absence of an "expert" login to actually search for the reason
of the growing diskspace is an ENTIRELY SILLY IDEA. I'm feeling like a
lobotomized ex-hacker at this "admin" shell...
Thanks,
Andre.
--
.sig making fun of Santa Claus Operation currently unavailable
-> Andre Beck +++ ABP-RIPE +++ IBH Prof. Dr. Horn GmbH, Dresden <-
________________________________________
_______
cisco-voip mailing list
cisco-voip@puck.nether.net
https://puck.nether.net/mailman/listinfo/cisco-voip
| |
| Andre Beck 2007-10-09, 7:11 am |
| Hi Ryan,
On Mon, Oct 08, 2007 at 09:57:33AM -0400, Ryan Ratliff wrote:
> Until you start getting RTMT alerts about the active partition being
> full don't worry about it. 93% utilization is pretty normal for a CM
> 5 box.
Ok ;)
Meanwhile we also see occasions where the used space actually dropped
down a bit and the partition with 5.1(3) is currently at 91% only. So
I'm relaxed about that, it just seemed odd.
> There were however some bugs that caused the active partition to fill
> up. CSCsk18010 is one of them, but only affects IBM servers and is
> fixed in 5.1(3).
This one is a HP DL380 with 146GB disks and we applied the patch to
bump up to 5.1(3) yesterday. So far all looks Ok, but it did so for
more than a week after bumping to 5.1(2) as well...
> I'd encourage you to open a case with TAC to troubleshoot Tomcat (the
> web server) and your h.323 issue. They definitely sound like
> problems but disk utilization is not the cause.
I'll have to open a case should the same situation ever arise again
after the upgrade, as there will be no other potential way out then.
Of course I hope 5.1(3) might have fixed whatever it was ;)
Thanks again,
Andre.
--
.sig making fun of Santa Claus Operation currently unavailable
-> Andre Beck +++ ABP-RIPE +++ IBH Prof. Dr. Horn GmbH, Dresden <-
| |
|
|
|
|
|