Menu
▾
▴
Re: [SL] Apache error running summary reports - reality checking script?
|
From: Michael H. <mh...@it...> - 2008-07-21 17:43:03
|
All valid points except that "top" shows CPU/RAM to be just fine
while the report is running, as is disk/swap.
This is an older machine, our last non-virtual system, actually,
so maybe we'll just go ahead and do the VM migration and see what
happens.
Thanks,
Michael
On Jul 21, 2008, at 8:59 AM, Paul Tammes wrote:
> Seeing FC4 where 9 is out, postgresql 7.4 where 8.x is out, could
> it be the
> hardware is old as well?
> I know, if it ain't broke and all that ;-)
>
> If the hardware is not up to speed for a larger number of months,
> memory
> overflow combined with swapping to (slow/old/olmost dead?) Harddisk
> comes to
> mind. The change in response between say an AMD Duron 800 and AMD
> 4450e
> dualcore is stunning.
> Both el cheapo cpu but frome different era's alltogether: in
> hardware 8
> years is a lifetime and a halve.
> Nice part: the power consumption of some modern cpu's like 4450e
> ore dulcore
> intels is stunningly lower.
> So that kind of investment earns itself in a few years time as well.
>
> (If that is too much of an investment, try adding some extra memory
> to the
> old box.
> SD-ram is still available in some antique shops, or second hand on
> computer
> user club trade fairs.
> DDR (sec) is more readilly available, and other than SD-ram (which
> is rare,
> therefore expensive) pretty cheap as well
>
> Hth
>
> 2008/7/21 Michael Hasse <mh...@it...>:
>
>> Further testing shows that it is indeed something timing out. The
>> reason the test company report ran okay was because there were hardly
>> any records. I am also able to run the problem reports okay provided
>> I choose only a month or two range. Having run the reports for
>> successive months of the entire year with no problems we can at least
>> get the taxes taken care of, (whew! :), and short of some bizarre
>> boundary error this should prove that the data is okay.
>>
>> But the question remains of what exactly it is that is timing
>> out...?
>>
>> In the process of testing this it looks like we actually get any
>> of a variety of errors, some of them pretty generic, e.g. plain old
>> Apache "Internal server error".
>> (William's suggestions earlier unfortunately did not pan out.)
>>
>> This is an FC4 box running:
>> Apache 2.0.54 w/mod_perl 2.0.1
>> Postgresql 7.4.13
>> Perl 5.8.6
>>
>> Any ideas?
>>
>>
>> Thanks!
>>
>> Michael
>>
>>
>> ================================
>> Based on the advice to check dates I've just been through every date
>> value of every record of every table, and I found a couple of bogus
>> dates, (you'd think bounds checking would catch '08-27-0200'), and
>> fixed them, but still no go.
>> I also checked to see if maybe the version upgrade had missed a
>> file due to permissions or something but all the .pl files are the
>> same date and perms, (except setup.pl is a few days newer but that's
>> to be expected and shouldn't matter).
>> What other things should we be checking?
>> Mismatches between invoices and line items, (e.g. line items
>> without invoices)?
>>
>> <rant>
>> I'm really surprised there are no scripts already written for
>> this. I actually was a rocket scientist, ahem, database programmer,
>> in a past life and every project I ever worked on had an "admin
>> toolkit" by the end of it with various queries to be sure nothing
>> bizarre had slipped through the cracks. This wasn't intentional, but
>> all db programmers end up doing some "pre-QA" of their own and the
>> scripts just grow naturally from that.
>> </rant>
>>
>> Anyway, again, suggestions would be greatly appreciated from
>> anyone who has dealt with anything like this before...?
>>
>>
>> Thanks!!!!
>>
>> Michael
>>
>>
>>
>> On Jul 3, 2008, at 4:06 AM, ja...@in... wrote:
>>
>>> A simple sanity check that I wish I had time to implement is to
>>> regularly (cron job) :
>>> Check each company in the database and hunt for any transactions
>>> with illogical (ie stupid) dates
>>> such as 07/03/2080 (transposed year digits)
>>> Send informative email to user who entered the offending
>>> transaction, and copy to sysadmin.
>>> Still hoping Dieter might implement something sensible in company
>>> defaults to solve this simple problem.
>>> Been mentioned before.
>>> Very interested in any other sanity checks.
>>> DJ
>>> On Thu 03/07/08 3:21 PM , Michael Hasse mh...@it... sent:
>>> Months later, I'm getting back to this, (prodded by tax
>>> deadlines of
>>>
>>> course! :)
>>> I did check the timeout directive and raised it to some
>>> ridiculous level to no avail, (and same duration of wait so I don't
>>>
>>> think that's it).
>>> Also don't think it's a record count issue as it was working
>>> fine
>>> right up a month or two before the upgrade and we don't do THAT
>>> much
>>>
>>> business.
>>> I also opened a terminal session to the host server and watched
>>>
>>> the stats as it was running and there were no issues with
>>> memory/swap/
>>> etc running out or anything like that.
>>> Somebody suggested checking another company in the system,
>>> (good
>>>
>>> idea, my apologies for not doing that before posting! Brain
>>> fade...).
>>> Unfortunately that company works just fine so I have to assume
>>> at
>>> this point that I have a corrupt record somewhere/somehow.
>>> So now the question is - are there any sort of scripts that can
>>>
>>> be run to do a sanity check? Has anybody had to deal with bad
>>> records and might have some tips to save a few hours of poring over
>>>
>>> the DB structure?
>>> Thanks!!!!!!!!!!!!
>>> Michael
>>> On Jan 8, 2008, at 6:01 AM, William McKee wrote:
>>>> On Mon, Jan 07, 2008 at 09:34:39PM -0800, Michael Hasse wrote:
>>>>> Just an update on this. Permissions are all fine and the reports
>>> run
>>>>> with no problems in our test company. Reading up a bit on the
>>> errors
>>>>> it looks like a simple timeout problem where Apache is giving up
>>> on
>>>>> the script before it finishes due to the amount of data we have
>>>>> slowing the script down. (System resources are okay so we're not
>>>
>>>>> running out of memory or anything).
>>>>
>>>> Interesting, I hadn't considered timeout error causing that
>>> message.
>>>>
>>>>
>>>>> Suggestions/advice are welcome and we'll keep plugging away
>>> at it
>>>>> and post back if we figure it out.
>>>>
>>>> Have you checked out the TimeOut directive[1]? I've seen some
>>> people
>>>> comment the TimeOut does not affect CGI scripts, but Apache
>>> security
>>>> tips[2] say the following:
>>>>
>>>> "As TimeOut is currently used for several different operations,
>>>
>>>> setting
>>>> it to a low value introduces problems with long running CGI
>>> scripts."
>>>>
>>>> I've seen others suggest using FastCGI. Keep us posted on your
>>>> progress.
>>>>
>>>>
>>>> Good luck,
>>>> William
>>>>
>>> --------------------------------------------------------------------
>>> --
>>> ---
>>> Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW!
>>> Studies have shown that voting for your favorite open source
>>> project,
>>> along with a healthy diet, reduces your potential for chronic
>>> lameness
>>> and boredom. Vote Now at http://www.sourceforge.net/community/cca08
>>> _______________________________________________
>>> sql-ledger-users mailing list
>>> sql...@li...
>>> https://lists.sourceforge.net/lists/listinfo/sql-ledger-users
>>
>>
>> ---------------------------------------------------------------------
>> ---
>> -
>> Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW!
>> Studies have shown that voting for your favorite open source project,
>> along with a healthy diet, reduces your potential for chronic
>> lameness
>> and boredom. Vote Now at http://www.sourceforge.net/community/cca08
>> _______________________________________________
>> sql-ledger-users mailing list
>> sql...@li...
>> https://lists.sourceforge.net/lists/listinfo/sql-ledger-users
>>
>> ---------------------------------------------------------------------
>> ----
>> This SF.Net email is sponsored by the Moblin Your Move Developer's
>> challenge
>> Build the coolest Linux based applications with Moblin SDK & win
>> great
>> prizes
>> Grand prize is a trip for two to an Open Source event anywhere in
>> the world
>> http://moblin-contest.org/redirect.php?banner_id=100&url=/
>> _______________________________________________
>> sql-ledger-users mailing list
>> sql...@li...
>> https://lists.sourceforge.net/lists/listinfo/sql-ledger-users
>>
>>
> ----------------------------------------------------------------------
> ---
> This SF.Net email is sponsored by the Moblin Your Move Developer's
> challenge
> Build the coolest Linux based applications with Moblin SDK & win
> great prizes
> Grand prize is a trip for two to an Open Source event anywhere in
> the world
> http://moblin-contest.org/redirect.php?banner_id=100&url=/
> _______________________________________________
> sql-ledger-users mailing list
> sql...@li...
> https://lists.sourceforge.net/lists/listinfo/sql-ledger-users
|
