On Wednesday 05 October 2005 23:34, Shang-Lin Chen wrote:
>Paddy Sreenivasan wrote:
>> Use ipcrm -m <id> to remove them and try amflush again.
>
>It worked! Thanks for your help. :-)
>
>Shang-Lin
Yes, but I think this one bears watching. Perhaps a miss-match in the
memory allocations methods amanda uses vs the Solaris methods is
causeing them to be left lying around. Not a Good Thing(TM).
Have cron run this "ipcs -m" sometime after amanda should be done, and
mail you the results every morning.
>> It appears taper is not cleaning up properly. tape_writer_side should
>> call detach_buffers() in case 'Q' to explicitly detach the shared
>> memory (might be required for Solaris)
>>
>> Paddy
>>
>>
>> On 10/5/05, *Shang-Lin Chen* <schen AT gps.caltech DOT edu
>> <mailto:schen AT gps.caltech DOT edu>> wrote:
>>
>> On Wed, 2005-10-05 at 16:39 -0700, Paddy Sreenivasan wrote:
>> > Maybe some other processes is using more shared memory or shared
>> > memory segments.
>> > Check output of ipcs (ipcs equivalent in Solaris)
>>
>> This is the output I'm getting from ipcs -m:
>> IPC status from <running system> as of Wed Oct 5 17:25:34 PDT
>> 2005 T ID KEY MODE OWNER GROUP
>> Shared Memory:
>> m 0 0x50000674 --rw-r--r-- root root
>> m 1 0 --rw------- amanda amanda
>> m 2702 0 --rw-rw-rw- webmgr staff
>> m 1203 0 --rw------- amanda amanda
>> m 104 0 --rw------- amanda amanda
>> m 5 0 --rw------- amanda amanda
>> m 106 0 --rw------- amanda amanda
>> m 207 0 --rw------- amanda amanda
>> m 8 0 --rw------- amanda amanda
>> m 109 0 --rw------- amanda amanda
>> m 210 0 --rw------- amanda amanda
>> m 511 0 --rw------- amanda amanda
>> m 312 0 --rw------- amanda amanda
>> m 1213 0 --rw------- amanda amanda
>> m 1014 0 --rw------- amanda amanda
>> m 115 0 --rw------- amanda amanda
>> m 716 0 --rw------- amanda amanda
>> m 17 0 --rw------- amanda amanda
>> m 118 0 --rw------- amanda amanda
>> m 319 0 --rw------- amanda amanda
>> m 620 0 --rw------- amanda amanda
>> m 221 0 --rw------- amanda amanda
>> m 622 0 --rw------- amanda amanda
>> m 523 0 --rw------- amanda amanda
>> m 224 0 --rw------- amanda amanda
>> m 125 0 --rw------- amanda amanda
>> m 126 0 --rw------- amanda amanda
>> m 227 0 --rw------- amanda amanda
>> m 128 0 --rw------- amanda amanda
>> m 1029 0 --rw------- amanda amanda
>> m 2330 0 --rw------- amanda amanda
>> m 231 0 --rw------- amanda amanda
>> m 1532 0 --rw------- amanda amanda
>> m 33 0 --rw------- amanda amanda
>> m 2034 0 --rw------- amanda amanda
>> m 235 0 --rw------- amanda amanda
>> m 436 0 --rw------- amanda amanda
>> m 1037 0 --rw------- amanda amanda
>> m 638 0 --rw------- amanda amanda
>> m 539 0 --rw------- amanda amanda
>> m 140 0 --rw------- amanda amanda
>> m 141 0 --rw------- amanda amanda
>> m 1542 0 --rw------- amanda amanda
>> m 643 0 --rw------- amanda amanda
>> m 144 0 --rw------- amanda amanda
>> m 345 0 --rw------- amanda amanda
>> m 1346 0 --rw------- amanda amanda
>> m 1447 0 --rw------- amanda amanda
>> m 848 0 --rw------- amanda amanda
>> m 1849 0 --rw------- amanda amanda
>> m 50 0 --rw------- amanda amanda
>> m 151 0 --rw------- amanda amanda
>> m 152 0 --rw------- amanda amanda
>> m 153 0 --rw------- amanda amanda
>> m 154 0 --rw------- amanda amanda
>> m 155 0 --rw------- amanda amanda
>> m 556 0 --rw------- amanda amanda
>> m 257 0 --rw------- amanda amanda
>> m 258 0 --rw------- amanda amanda
>> m 659 0 --rw------- amanda amanda
>> m 1160 0 --rw------- amanda amanda
>> m 261 0 --rw------- amanda amanda
>> m 262 0 --rw------- amanda amanda
>> m 363 0 --rw------- amanda amanda
>> m 64 0 --rw------- amanda amanda
>> m 265 0 --rw------- amanda amanda
>> m 66 0 --rw------- amanda amanda
>> m 267 0 --rw------- amanda amanda
>> m 468 0 --rw------- amanda amanda
>> m 769 0 --rw------- amanda amanda
>> m 170 0 --rw------- amanda amanda
>> m 171 0 --rw------- amanda amanda
>> m 172 0 --rw------- amanda amanda
>> m 973 0 --rw------- amanda amanda
>> m 1074 0 --rw------- amanda amanda
>> m 1275 0 --rw------- amanda amanda
>> m 4176 0 --rw------- amanda amanda
>> m 577 0 --rw------- amanda amanda
>> m 878 0 --rw------- amanda amanda
>> m 479 0 --rw------- amanda amanda
>> m 80 0 --rw------- amanda amanda
>> m 381 0 --rw------- amanda amanda
>> m 382 0 --rw------- amanda amanda
>> m 183 0 --rw------- amanda amanda
>> m 84 0 --rw------- amanda amanda
>> m 185 0 --rw------- amanda amanda
>> m 1186 0 --rw------- amanda amanda
>> m 1887 0 --rw------- amanda amanda
>> m 4688 0 --rw------- amanda amanda
>> m 689 0 --rw------- amanda amanda
>> m 190 0 --rw------- amanda amanda
>> m 191 0 --rw------- amanda amanda
>> m 692 0 --rw------- amanda amanda
>> m 693 0 --rw------- amanda amanda
>> m 1094 0 --rw------- amanda amanda
>> m 195 0 --rw------- amanda amanda
>> m 1496 0 --rw------- amanda amanda
>> m 1397 0 --rw------- amanda amanda
>> m 298 0 --rw------- amanda amanda
>> m 499 0 --rw------- amanda amanda
>>
>> However, when I did
>> ps -ef | grep amanda
>> I couldn't find any processes running as amanda. I also tried
>> running amflush using the next tape in the cycle (I couldn't find any
>> spare tapes that worked with the drive) and got the same error about
>> no space
>> left on the device.
>>
>> Shang-Lin
>>
>> > Paddy
>> >
>> > On 10/5/05, Shang-Lin Chen <schen AT gps.caltech DOT edu
>>
>> <mailto:schen AT gps.caltech DOT edu>> wrote:
>> > The Amanda server is running Solaris, so it doesn't
>> > have /proc. But I
>> > checked /etc/system and found this:
>> > set shmsys:shminfo_shmmax=268435456
>> > set shmsys:shminfo_shmmin=1
>> > set shmsys:shminfo_shmmni=100
>> > set shmsys:shminfo_shmseg=10
>> > set semsys:seminfo_semmns=70
>> > set semsys:seminfo_semmni=100
>> > set semsys:seminfo_semmsl=70
>> > set semsys:seminfo_semopm=100
>> > set semsys:seminfo_semvmx=32767
>> > forceload: sys/shmsys
>> >
>> > If these settings worked before, how come amflush is
>> > failing now?
>> >
>> > Thanks,
>> > Shang-Lin
>> >
>> > On Wed, 2005-10-05 at 14:54 -0700, Paddy Sreenivasan
wrote:
>> > > Can you check the number of shared memory segments and
>> >
>> > shared memory
>> >
>> > > allocated in the system? You may have to adjust the
>>
>> values
>>
>> > of
>> >
>> > > shmall and shmmni (cat /proc/sys/kernel/shmmni)
>> > >
>> > > Paddy
>> > >
>> > > On 10/5/05, Shang-Lin Chen < schen AT gps.caltech DOT edu
>>
>> <mailto:schen AT gps.caltech DOT edu>> wrote:
>> > > I added
>> > > tapebufs 50
>> > > to amanda.conf after the labelstr line, and
>>
>> this is
>>
>> > the output
>> >
>> > > I'm
>> > > getting now:
>> > >
>> > >
>> > > FAILURE AND STRANGE DUMP SUMMARY:
>> > > taper: FATAL shmget: (1638600 tapebufs) No
>> > > space
>> >
>> > left on
>> >
>> > > device
>> > > amflush: ERROR Cannot flush without tape.
>> > > Try
>> >
>> > again.
>> >
>> > > Shang-Lin
>> > >
>> > > On Wed, 2005-10-05 at 17:02 -0400, Scott R.
>> > > Burns
>> >
>> > wrote:
>> > > > Try
>> > > >
>> > > > tapebufs 50
>> > > >
>> > > > In your amanda.conf file
>> > > >
>> > > > Scott...
>> > > >
>> > > > Scott R. Burns
>> > > > NETCON Technologies Incorporated
>> > > > Suite 135 - 4474 Blakie Road
>> > > > London, Ontario, Canada
>> > > > N6L 1G7
>> > > > Voice: +1.519.652.0401
>> > > > Fax: +1.519.652.9275
>> > > > Web: www.netcontech.com
>>
>> <http://www.netcontech.com>
>>
>> > > > -----Original Message-----
>> > > > From: Shang-Lin Chen [mailto:
>> >
>> > schen AT gps.caltech DOT edu <mailto:schen AT gps.caltech DOT
>> edu>]
>> >
>> > > > Sent: Wednesday, October 05, 2005 4:21 PM
>> > > > To: Paul Bijnens
>> > > > Subject: Re: no space left on device during
>> >
>> > amflush
>> >
>> > > > > Shang-Lin Chen wrote:
>> > > > > > taper: FATAL shmget: (655440 tapebufs)
>> > > > > > No
>> >
>> > space left
>> >
>> > > on device
>> > >
>> > > > > The default value for tapebufs is 20
>>
>> (buffers of
>>
>> > 32K bytes
>> >
>> > > each).
>> > >
>> > > > > You probably got some weird value in
>>
>> amanda.conf
>>
>> > for that
>> >
>> > > parameter.
>> > >
>> > > > > The above value results if over 20 GByte
>>
>> of RAM.
>>
>> > > > My amanda.conf file doesn't set anything for
>> >
>> > tapebufs, and
>> >
>> > > I've never
>> > >
>> > > > had this problem before. How do I change the
>> >
>> > tapebufs
>> >
>> > > setting?
>> > >
>> > > > Thanks,
>> > > > Shang-Lin
--
Cheers, Gene
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.35% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.
|
