are RESERVED, but the mount
request for the second MP doesn't get dispatched until the first is
complete. This serialization can waste a lot of wall-clock time.
Does anyone on the list know a reason for this to be so?It seems an
odd change.
- Allen S. Rout
On 01/28/2014 11:32 PM, Prather, Wanda wrote:
Gotchas?
Likes?
Hi, Wanda. Long time no chat.
Is this blast from the past still an interesting topic to you?
I'm doing UCS.
- Allen S. Rout
nfig for [database] using logprimary [n/512]
Is anyone using logsecondary to manage overflow space?
- Allen S. Rout
S. Rout
are recorded in an addendum to the file, on the
storage pool. This means that, if the extended permissions change, the
whole file must be re-backed-up.
So, it's not really a mystery why an heritable change at the root of a
FS would cause a resending of the whole FS.
- Allen S. Rout
bottleneck is in fact on one of the EMC boxen.
This will require periodic attention; in your shoes, I'd write a
script to do the analysis too, which would have as its output a dump
of N lists of nodes, tweaked according to your evolving comprehension
of the bottleneck.
- Allen S. Rout
is the bottleneck, usually.
And if one mount point ran 11.5 hours, then I'm betting your peak incr
count is pretty high.
- Allen S. Rout
be RESOURCEUTILIZATION
whatever) are running at the same time on your proxy box? If that's 31,
you're probably swapping your butt off.
+ How many filespace walking threads are running simultaneously on your
various back-end storage systems?
I think you'll need to analyze at a tighter than server level.
- Allen S
a payload ca. 132M. For empty. As in, all it's
got to 'incr' about is the activity log lines reporting the completion
of the previous backup.
Any notions?
- Allen S. Rout
On 08/19/2013 05:07 PM, Remco Post wrote:
Hi Allen,
Which version of TSM are you using?
6.3.
- Allen S. Rout
=STGPOOL_NEW
Once you've got all of NODE1's data out of TEMP, you can set TEMP's
NEXTSTG to be STGPOOL_OLD, and set highmig and lowmig to zero, and wait
for it to drain.
- Allen S. Rout
be a challenge.
I'd be really surprised if the _node_ count, per se, was relevant.
I've got some customers who have dozens of nodes on a single server,
because they want to be able to move them independantly.
You're not using 31 different DSMCAD processes, are you? :)
- Allen S. Rout
proxies I've run into, the memory footprint necessary to
keep the whole manifest in core was bigger than folks wanted to spend.
- Allen S. Rout
space delete has been running for several hours, and
I need to do more deletes (this node and others) But don't want to risk
trouble.
I suggest running a macro or script, with the successive deletes, WAIT=YES.
- Allen S. Rout
our Oracle backups off of TSM entirely.
On Tue, Jul 23, 2013 at 3:12 PM, Allen S. Rout a...@ufl.edu wrote:
We're not using boost; our primary use for the DD is for Oracle
backups, and our DBAs are far more interested in the conventional
filesystem user interface than they are in the network
level and at the storage array level. I've
heard several rumors that the Data Domain can dedupe even deduped
e.g. VEEAM backups, with very good ratios. I'm going to try a
similar theory with the DD and TSM-deduped stgpools.
- Allen S. Rout
.
- Allen S. Rout
likely to use it: to a man in the desert, brackish
swampwater will do. But I can wish for some good old NYC tap, can't I?
- Allen S. Rout
irritating.
Because if I can, then I'm going to have an infarction right now.
They've been talking about external authentication for at least 12 years.
- Allen S. Rout
recommended
against.
- Allen S. Rout
not just for
[product X].
I'm musing about sending my DB backups straight to a Data Domain, and
syncing that offsite. It'd take away one level of indirection from my
DR planning.
- Allen S. Rout
is a critical performance enhancement for DR, and substantial for
normal access, too. Well tuned, he thinks 95% reduction in the SQL
database size is readily achievable. At that rate, you can fit the
remaining 5% on SSD...
- Allen S. Rout
problem. If you've got 20M files, that's still a big problem, but at
least you've separated the conjoined twins.
- Allen S. Rout
successfully deleted.
tsm: TSMPOK_SERVER1q volh t=dbs
Date/Time: 05/24/2013 05:11:32
5 days ago?
- Allen S. Rout
) domain. These folks
can register new nodes, delete their data, etc.
We assign charging authority by-domain and by-node.
- Allen S. Rout
. :)
- Allen S. Rout
.
- Allen S. Rout
approached that was to
dsmadmc -comma -dataonly=yes
and do my formatting in my scripting engine.
- Allen S. Rout
On 04/16/2013 04:58 PM, Alex Paschal wrote:
Back in the day, it almost always used to be a duplex mismatch. Get off
my lawn! :-)
That answers the description of 'transmitted relatively slowly', now,
doesn't it? :)
- Allen S. Rout
- Throwing down with ADSM 2.1, yo yo. ;)
in the past,
which evaluates collocation for me, but I think it's poorly aligned with
your needs, Grant: I select candidates for inclusion into collocation
groups, and then once that's done I MOVE DATA to vacate tapes which are
low-utilization and have a candidate collocated home.
- Allen S. Rout
perfect' or such,
no alert is required when we spill over.
- Allen S. Rout
On 04/16/2013 08:59 AM, Zoltan Forray wrote:
The active session is only up to 222GB so it isn't volume - just one long
transaction, I guess?
(reliatively) Big files, transmitted (relatively) slowly, are my most
common log-pin culprit.
- Allen S. Rout
On 04/16/2013 09:36 AM, Zoltan Forray wrote:
On Tue, Apr 16, 2013 at 9:22 AM, Allen S. Rout a...@ufl.edu wrote:
On 04/16/2013 08:59 AM, Zoltan Forray wrote:
The active session is only up to 222GB so it isn't volume - just one long
transaction, I guess?
(reliatively) Big files, transmitted
, and I want to
double-check that this is the case at this transition, too. Or am I
insane and remembering incorrectly?
Any war stories about drive hardware updates?
- Allen S. Rout
a human
pace.The 3584s just... lose me. They're keen.
- Allen S. Rout
has worked in the past. The most recent
database restore I did from virtual volumes was with 6.2.2 - could someone
from development also comment regarding this new restriction please.
Eek!
- Allen S. Rout
you manage it from the human
side. :)
- Allen S. Rout
don't really have a clear idea how long that has
been getting in the way of normal operations.
- Allen S. Rout
and the behavior hasn't yet repeated.
- Allen S. Rout
On 01/15/2013 12:46 PM, Prather, Wanda wrote:
Is it V6, and could it be DB2 processes doing the I/O instead of the TSM
server itself?
No; 5.5.
- Allen S. Rout
deployment, however, expects some 300-400TB of
capacity.
- Allen S. Rout
to the desktop needs to be thinking about
local imperatives.
I like central homedirs as a solution to that problem, but that's kind
of not the question you asked. :)
- Allen S. Rout
it's not an
(increasingly long list of) manual operations.
- Allen S. Rout
.
+ Disaster-recovery needs addressed by replication
+ Longer-term backup needs addressed by incrementals to TSM, with
retention set to keep enough data. e.g.
Monthlies for 2 years
verexists: 26
verdeleted: 26
retextra: 800
retonly: 800
and do incrementals monthly.
- Allen S. Rout
volumes, and reconcile them.
- Allen S. Rout
a filesystem export/import
to a different node name.
- Allen S. Rout
on proxy-08.
There, you've got some issues; the fully deployed solution to that is
to associate each filespace with its own nodename, and then you can move
nodes _and_ filespaces around on proxy hardware to your hearts' content.
- Allen S. Rout
a separate filesystem/'disk' per LUN, and I'd always
put your stuff in a subdirectory on the volume. That way you can, at
need, make a
/volumes/tablespace02/tmp
parallel to
/volumes/tablespace02/containers
or such.
- Allen S. Rout
. I'd be afraid that it would mostly work,
as in not show problems for A while, and then leave you up a
semiliquid creek.
- Allen S. Rout
implicit in
e.g. a DOMAIN statement, and you have to do something like
def sched [domain] [name] [ action=incr ] OBJ=/nasfs
OPT='-snapshootroot=/nasfs/.snapshot/nightly.0'
- Allen S. Rout
in a temporary script, and then run that as a COMMAND schedule
instead of an INCREMENTAL sched.
- Allen S. Rout
collocate by filespace. :)
- Allen S. Rout
with occupation on a volume.
I'm moving architecture _and_ version (AIX to Linux, sigh) so I figure
since I'll be recreating my whole infrastructure one step at a time,
it's a GREAT time to rethink everything.
- Allen S. Rout
or VERDELETED. Interesting.
- Allen S. Rout
contention.
- Allen S. Rout
no panacea, to be
sure. But...
- Allen S. Rout
On 07/23/2012 06:03 PM, Kevin Kettner wrote:
http://xkcd.com/936/
I think this reference rebuts your desire, rather than supporting it.
In fact, I was already looking for the comic in order to rebut, before I
got to the end of your message. :)
- Allen S. Rout
I'm getting a remarkably official-looking This page is no longer
available from the passport advantage customer login links scattered
all over IBM web space.
You folks seeing the same? The page tastes like a 'removed this
feature' sort of page, not like a D'oh, app busted! page.
- Allen S
-- completely unable to parse it
when I click on the 'Customer Sign In' link on the PAO site. FF and
chrome both.
Oh, well.
- Allen S. Rout
do the same thing for
mount behavior.
- Allen S. Rout
on this is that you _want_ to wipe and rebuild when
you reinstall. You ought to have your configuration control well
enough in place that this doesn't occupy a lot of human time. And if
it's not quite there, each burn-and-restart iteration is another
opportunity to improve the automation.
- Allen S. Rout
to
the new location, assuming it is the same path on the new server.
Works for the final exercise, but not for testing.
- Allen S. Rout
: 3.9
Original Bytes:6,995,806,860
Globally Compressed:6,453,299,125
Locally Compressed:1,774,055,288
Meta-data: 19,701,904
so that equates to the volume level if you use a collocated FILE devclass.
- Allen S. Rout
of that, but it's important to
keep in mind that they sum.
- Allen S. Rout
with.
- Allen S. Rout
of machines can rationally be excluded from SYSTEM STATE
backups: Offhand, I think their DR plans don't include BMR from my TSM.
But that's short-term thinking.
What are you-all doing about this? Increasing number of DB incrs
between fulls? Something else?
- Allen S. Rout
of notes,
specifically on Oracle questions. I'm not chasing
If your response to this is No, but I'd like to FIND one, then please
reply to me off the list; if sufficient numbers do so, I'll gin up a
list here.
- Allen S. Rout
My apologies: I try to avoid writing to the readership of a list, but
this is just neat.
http://en.wikipedia.org/wiki/Linear_Tape_File_System
- Allen S. Rout
the assemblage Just In
Case.
- Allen S. Rout
On 03/06/2012 12:21 PM, Karel Bos wrote:
Ofc there is, just restore the tsm database to a point be4 the data being
inactivated :)
Heh. Yeah, I'd thought of that. :)
- Allen S. Rout
didn't know that. Unfortunately, that wouldn't let them
skip the backup. Thanks, I didn't think I'd missed something, but it's
nice to be able to compare notes with gurus.
- Allen S. Rout
on the management
costs of running multiple environments.
What he said. :)
- Allen S. Rout
servers. I'd
already warned folks that they were going out of support, but...
I expect that backlevel clients will work in most normal environments;
simple file incrementals et. alia.. but they should generate at least a
touch of organizational angst.
- Allen S. Rout
backups for user experience. Again, so long as you don't compress or
encrypt the streams, some of the data (maybe even much?) should
dedupe neatly.
- Allen S. Rout
900/s and 1300/s.
AIX, 5.5, on an EMC VNX with Fast Cache (SSD caching).
- Allen S. Rout
, but it moves you to a totally different recovery
mode.
Our thinking abount isilon backups includes a remote Near-line unit,
that has relatively large, slow, cheap drives. It would be a
replication target for several other units, giving it good efficiency
of scale.
- Allen S. Rout
On 02/06/2012 05:33 PM, Paul_Dudley wrote:
Re your line None of them has data from more than one client
Has collocation been turned on by accident?
No, it has
The tapes are collocated by node. There are about 20 server nodes which
write to it.
been turned on by design.
- Allen S. Rout
plans so far, we're thinking 'replicate and snap' for DR
purposes. Longer term archives aren't currently under consideration,
but pleasantly they don't have the same frequency issues. If it takes
me a week to do an incr once a year, I'm less concerned.
- Allen S. Rout
So, I've got an offsite machine which exists to accept remote virtual
volumes. For years, now, the filling volumes have behaved in a way I
thought I understood.
The tapes are collocated by node. There are about 20 server nodes which
write to it.
My number of filling volumes has rattled around
of information; they might or might not be useful for what you're after.
- Allen S. Rout
On 01/20/2012 03:50 PM, Vandeventer, Harold [BS] wrote:
I'm confident they're empty; just trying to prove it before I delete.
CONTENT isn't providing anything, and I can display content of volumes in other
pools.
Are they PENDING?
- Allen S. Rout
due to changes in TSM 6.x, or has it
always been this way for SCSI-style (smcX) libraries?
I've never had this issue with my 3494 library and any TSM version
from ADSM 2.1 through TSM 5.5.
Me too; 3494 and 3584 both, I've not needed to restart TSM to begin
accessing new devices.
- Allen S. Rout
glmail01 glmail02 whole list of servers; \
tape-paths -server $server
which gives output you can just paste into your library manager.
- Allen S. Rout
#!/usr/local/bin/perl
use FileHandle;
use Data::Dumper;
my $verbose =1;
my $in = new FileHandle;
my $drives = {};
my $out = {};
$in-open(/var
this could very well be my problem. Thanks!
Shame I can't fix it for extant data. I might not be able to fix it at
all: if the content managment system sends just a page, or just a
packet of metadata, the txngroupmax will be irrelevant: transaction
size = 1. :P
- Allen S. Rout
',
I think 'copystgpool=[yadda]'. But nonblock is specifically allowed for
copystgpool sources. I don't use copystgpool, but it's of interest
nonetheless.
- Allen S. Rout
with a
reason a full 5G volume would have an ESTCAP of 1.3G with no reclaimable
space.
- Allen S. Rout
On 01/03/2012 03:14 PM, Richard Rhodes wrote:
Are the volumes all status FULL?
Hmm, I'd thought I'd said that. Yes, they're all full.
- Allen S. Rout
is
decades more complete, but most of that will be irrelevant for TSM
scripting.
[...] I can run it on either platform.
We play both kinds: Country _and_ Western.
- Allen S. Rout
for incrementals.
really the trade-off here is between equipment and expertise. It's
rational to choose the former, but the result will be that you're
doing things more expensively than is necessary.
- Allen S. Rout
has
had any early-adopter issues with it.
- Allen S. Rout
, whatever that means in your
environment. Several folks here (including me) have posted the
scripts they use to do their analysis and modifications, so it need
not be a major time sink.
- Allen S. Rout
into another stg. I
know that dedupe is per-pool, but I seem to recall that it works to
move from a dedupe pool to another dedupe pool?
- Allen S. Rout
.
If they are dealing with that in some different way...
- Allen S. Rout
a specific mention of virtual volumes. I ass-u-me that
it's similar to the library manager situation: i.e. a 6.2 server target
can interact with sources down to 5.5; but I don't seem to see
specific confirmation of that.
Anyone have feedback on that?
- Allen S. Rout
- Finally planning The Move.
the last time you ran an incremental of the
machine, it will stay that way, and you can expire inventory as often
as you like. Only inactive file versions will be eventually pared
away.
- Allen S. Rout
On 10/18/2011 10:48 AM, Paul Zarnowski wrote:
Allen,
[... cornucopia ...]
HTH.
That's fantastic. Do you chance to recall a first approximation of what
your DB sizes were when you left V5?
- Allen S. Rout
On 10/14/2011 03:54 PM, Robert Clark wrote:
The symposium example seems a classic case of selection bias.
:) Zing!
I agree. But it is at least a selection pool more like the membership
of this list than like the general population.
- Allen S. Rout
experiences w.r.t. memory footprint?
It's my major planning unknown; I'm musing about instance count
vs. size.
- Allen S. Rout
terabytes large, but
looks small on a 'q filespace').
5: Do (elementary) statistics on the results.
- Allen S. Rout
you.
For the change rate analysis, you're correct: 2) is the utilized
capacity, CAPACITY / PCT_UTIL.
- Allen S. Rout
1 - 100 of 696 matches
Mail list logo