Failed. Service unavailable when attempting test daily call please help!

[goldendel]

Hi Peter,

I've been getting "Failed. Service unavailable" again for the past few days.

Attached are the networking howto test results, listing of my oztivo.conf and my tclient.log

My Tivo is directly connected to a Netcomm NB12WD via a Netgear ethernet over power adapter. Can you check if there is anything obvious I am doing wrong?

Many thanks in advance for your help,

Tim

[petestrash]

Your networking and settings seem to fine.

Looking at the log it appears you may have some data corruption on your disk.

How old is the Drive you are using, and have you noticed any Audio or Video Glitches during playback?

Peter.

[goldendel]

Hi Peter,

I bought the disk on 5/08/2006 according to my records. I haven't seen any audio\video glitches so far. I think the disk is 34% full. I can delete some stuff, would that help?

Many thanks,

Tim

[petestrash]

Deleting shows should not make any difference.

We can try and have the TiVo fix any database and file system issues. If it does not work, re-imaging may be the only alternative.

Also running the commands may stop your TiVo from working if there is a lot of corruption, which means that you will have no choice but to re-image. This only happens in a small number of cases, but it can happen.

So I suggest you watch or transfer your important shows first.

There are 2 area to check. These commands first try to fix the internal database.

[TiVo [p0] ~]# chmod 777 /var/tmp//fsmem.mpkey
[TiVo [p0] ~]# rm /var/tmp//fsmem.mpkey
[TiVo [p0] ~]# mfscheck -fix

This can take many hours to complete.

Let me know the results if any, then we will move to the filesystem check.

Peter.

[goldendel]

It took a while to backup shows\settings. I ran the commands, but the scan failed, so I re-ran it with only 1 trailing slash and got the following:


[TiVo [p1] /var/tmp]# chmod 777 /var/tmp/fsmem.mpkey
[TiVo [p1] /var/tmp]#
[TiVo [p1] /var/tmp]# rm /var/tmp/fsmem.mpkey
[TiVo [p1] /var/tmp]#
[TiVo [p1] /var/tmp]# mfscheck -fix
mfscheck scan begins
Checking reference counts
All reference counts are OK.
mfscheck scan ends
mfscheck: 0 fatal errors, 0 severe errors, 0 warnings.

You must manually restart the EventSwitcher
[TiVo [p1] /var/tmp]#

It ran in under a couple of minutes.

Tim

[petestrash]

Ok, reboot the TiVo and now try

[TiVo [p0] ~]# fsfix

Peter.

[goldendel]

Hi Peter,

I ran fsfix twice (1st time my PC crashed due to an unrelated event, taking the telnet with it). Here is the output:

fsfix
fsfix: mounted MFS volume, starting consistency checks.
I-Nodes:
I-Node table size is 131072 entries for 102400 active nodes max.
FsId high-water mark is 0xf8ba
Pass 1 - scan and analyze
reconstructing zone buddymaps
synchronizing...
volume marked as OK
scanned 8272 files, covering 3163 extents
29072 application pages in use
599031808 media pages in use
Inode table collision details
0 hash collisions in node table.
Allocation Details
6224/7575 application inline files, 82.16%
25.50% wasted space (0 bytes) in normal application region
48.7% wasted space (0 bytes) in inlined application region
697 media files
1.54% wasted space (0 bytes) in media region
Fragmentation:
average extents/file is 0.38
worst extents/file is 9
expected extents/file (approx) is 3
Application region:
zone contains 785168 pages: 29072 allocated; 756096 free
inode allocations match allocated page count
buddy-map is internally consistent, 756096 pages marked free
Media region:
zone contains 2097152 pages: 2064384 allocated; 32768 free
inode allocations match allocated page count
buddy-map is internally consistent, 32768 pages marked free
zone contains 620625920 pages: 596967424 allocated; 23658496 free
inode allocations match allocated page count
buddy-map is internally consistent, 23658496 pages marked free
fsfix: 0 fatal errors, 0 warnings.

I'm not sure if it is related, but if I try to start ms_ftp, I get the error:

[TiVo [p2] ~]# /hack/mfs_ftp/mfs_ftp.tcl
No EventSwitcher active
while executing
"dbopen"
(file "/hack/mfs_ftp/mfs_ftp.tcl" line 1531)

Tim

[goldendel]

Hi Peter,

I noticed the Tivo was frozen, and realised it did not get a reboot after the mfscheck. I rebooted, re-ran mfs check:

[TiVo [p0] ~]# chmod 777 /var/tmp//fsmem.mpkey
[TiVo [p0] ~]#
[TiVo [p0] ~]# rm /var/tmp//fsmem.mpkey
[TiVo [p0] ~]#
[TiVo [p0] ~]# mfscheck -fix
mfscheck scan begins
Checking reference counts
All reference counts are OK.
mfscheck scan ends
mfscheck: 0 fatal errors, 0 severe errors, 0 warnings.

You must manually restart the EventSwitcher

The Tivo again froze, so I rebooted it and ran fsfix:

[TiVo [p0] ~]# fsfix
Couldn't CreateSharedMemoryFile() /var/tmp//fsmem.mpkey -- check permissions
Error in FsInit: 0x11001
[TiVo [p0] ~]#
[TiVo [p0] ~]# chmod 777 /var/tmp//fsmem.mpkey
[TiVo [p0] ~]#
[TiVo [p0] ~]# rm /var/tmp//fsmem.mpkey
[TiVo [p0] ~]#
[TiVo [p0] ~]# fsfix
fsfix: mounted MFS volume, starting consistency checks.
I-Nodes:
I-Node table size is 131072 entries for 102400 active nodes max.
FsId high-water mark is 0xf8f2
Pass 1 - scan and analyze
reconstructing zone buddymaps
synchronizing...
volume marked as OK
scanned 8060 files, covering 3116 extents
28944 application pages in use
596475904 media pages in use
Inode table collision details
0 hash collisions in node table.
Allocation Details
6033/7369 application inline files, 81.86%
25.32% wasted space (0 bytes) in normal application region
48.25% wasted space (0 bytes) in inlined application region
691 media files
1.12% wasted space (0 bytes) in media region
Fragmentation:
average extents/file is 0.38
worst extents/file is 9
expected extents/file (approx) is 3
Application region:
zone contains 785168 pages: 28944 allocated; 756224 free
inode allocations match allocated page count
buddy-map is internally consistent, 756224 pages marked free
Media region:
zone contains 2097152 pages: 2064384 allocated; 32768 free
inode allocations match allocated page count
buddy-map is internally consistent, 32768 pages marked free
zone contains 620625920 pages: 594411520 allocated; 26214400 free
inode allocations match allocated page count
buddy-map is internally consistent, 26214400 pages marked free
fsfix: 0 fatal errors, 0 warnings.

I hope this helps. By the way, after another reboot I can now start ms_ftp.

Tim

[Darren King]

If it was me I would now be pulling the drive and checking it on a PC with Seatools (available from www.seagate.com under "support" then "downloads") running the full drive scan.

I bought the disk on 5/08/2006 according to my records.

That's almost three years. In a TiVo that gets constant use I usually expect to see problems after 18-24 months.

[petestrash]

Both of these processes kill the event switcher and require a reboot to get the system running again.

But they did not find anything wrong in the structure.

If you have a spare drive I would suggested loading the latest image onto it and see if the problem goes away.

Peter.