[gpfsug-discuss] CES node slow to respond

Fri Mar 24 18:19:52 GMT 2017

changes in ganesha management code were made  in April 2016 to reduce the
need for high maxfilestocache value, the ganesha daemon adjusts it allowed
file cache by reading the maxfilestocache value and then reducing its
allowed NOFILE value . the code shipped with 4.2.2 release.
you want a high maxfilestocache value on CES nodes for various reasons.
just FYI we have some customers who have MFTC set to 15 Million :-)

sven

On Fri, Mar 24, 2017 at 11:13 AM Bryan Banister <bbanister at jumptrading.com>
wrote:

> Hi Vipul,
>
>
>
> Hmm… interesting.  We have dedicated systems running CES and nothing else,
> so the only  thing opening files on GPFS is ganesha.  IBM Support
> recommended we massively increase the maxFilesToCache to fix the
> performance issues we were having.  I could try to reproduce the problem to
> investigate further, but the impact to users would not be appreciated.
>
>
>
> Setting such a high value for maxFilesToCache has direct impact on the
> token manager so fixing this would be very good.
>
>
>
> Perhaps IBM could take a closer look at this without our involvement?
>
> -Bryan
>
>
>
> *From:* Vipul Paul [mailto:vpaul at us.ibm.com] *On Behalf Of *IBM Spectrum
> Scale
> *Sent:* Friday, March 24, 2017 12:18 PM
> *To:* gpfsug main discussion list <gpfsug-discuss at spectrumscale.org>;
> Bryan Banister <bbanister at jumptrading.com>
> *Subject:* Re: CES node slow to respond
>
>
>
> Hi Bryan,
>
> Making sure Malahal's reply was received by the user group.
>
> >> Then we noticed that the CES host had 5.4 million files open
>
> This is technically not possible with ganesha alone. A process can only
> open 1 million files on RHEL distro. Either we have leaks in kernel or some
> other processes contributing to this.
>
> Ganesha does keep NFSv3 files open and keep open for performance. It
> doesn't have a good logic to close them after some inactivity. It does
> close them if the number is close to max open files which is configurable.
>
> PS: kNFS does open, read/write, and then close.  No caching in older
> versions. They did have a feature in a recent code to cache open files in
> NFSv3 as well.
>
> Regards, The Spectrum Scale (GPFS) team
>
>
> ------------------------------------------------------------------------------------------------------------------
> If you feel that your question can benefit other users of  Spectrum Scale
> (GPFS), then please post it to the public IBM developerWroks Forum at
> https://www.ibm.com/developerworks/community/forums/html/forum?id=11111111-0000-0000-0000-000000000479.
>
>
> If your query concerns a potential software error in Spectrum Scale (GPFS)
> and you have an IBM software maintenance contract please contact
> 1-800-237-5511 <(800)%20237-5511> in the United States or your local IBM
> Service Center in other countries.
>
> The forum is informally monitored as time permits and should not be used
> for priority messages to the Spectrum Scale (GPFS) team.
>
>
>
> From:        Bryan Banister <bbanister at jumptrading.com>
> To:        gpfsug main discussion list <gpfsug-discuss at spectrumscale.org>
> Date:        03/23/2017 08:27 AM
> Subject:        Re: [gpfsug-discuss] CES node slow to respond
> Sent by:        gpfsug-discuss-bounces at spectrumscale.org
> ------------------------------
>
>
>
>
> Anybody from IBM willing/able to give us some explanation of why ganesha
> is holding open so many files?  Is this expected/needed/etc?
>
> Or do we have to open a PMR to get some kind of explanation?
> -B
>
> -----Original Message-----
> From: gpfsug-discuss-bounces at spectrumscale.org [
> mailto:gpfsug-discuss-bounces at spectrumscale.org
> <gpfsug-discuss-bounces at spectrumscale.org>] On Behalf Of Matt Weil
> Sent: Thursday, March 23, 2017 10:24 AM
> To: gpfsug-discuss at spectrumscale.org
> Subject: Re: [gpfsug-discuss] CES node slow to respond
>
> FYI all
>
> We also ran into this after bumping maxFilesToCache.
>
> Mar 22 13:02:37 ces1 ntpd[1191]: ./../lib/isc/unix/ifiter_ioctl.c:348:
> unexpected error:
> Mar 22 13:02:37 ces1 ntpd[1191]: making interface scan socket: Too many
> open files in system
>
> fix
>
> sysctl -w fs.file-max=1000000
>
>
> On 3/22/17 12:01 PM, Bryan Banister wrote:
> > We had a similar issue and also were instructed by IBM Support to
> increase the maxFilesToCache to an insane value... Basically when the file
> cache gets full then the host will spend all of its cycles looking for a
> file to evict every time a new file is opened...  baaaah.
> >
> > Not sure why Ganesha has to keep so many files open... I can't believe
> our NFS clients actually keep that many open.  cNFS never needed this.
> > -Bryan
> >
> > -----Original Message-----
> > From: gpfsug-discuss-bounces at spectrumscale.org [
> mailto:gpfsug-discuss-bounces at spectrumscale.org
> <gpfsug-discuss-bounces at spectrumscale.org>] On Behalf Of Matt Weil
> > Sent: Wednesday, March 22, 2017 11:43 AM
> > To: gpfsug main discussion list <gpfsug-discuss at spectrumscale.org>
> > Subject: [gpfsug-discuss] CES node slow to respond
> >
> > All,
> >
> > We had an indecent yesterday where one of our CES nodes slowed to a
> > crawl.  GPFS waiters showed pre fetch threads going after inodes.
> > iohist also showed lots of inode fetching.  Then we noticed that the CES
> > host had 5.4 million files open.
> >
> > The change I made was to set maxStatCache=DEFAULT because it is linux.
> > And set maxFilesToCache=10000000 it was set to 500000.  Then restarted
> GPFS.
> >
> > Is there something else we should change as well.
> >
> > Thanks
> >
> > Matt
> >
> >
> > ________________________________
> > The materials in this message are private and may contain Protected
> Healthcare Information or other information of a sensitive nature. If you
> are not the intended recipient, be advised that any unauthorized use,
> disclosure, copying or the taking of any action in reliance on the contents
> of this information is strictly prohibited. If you have received this email
> in error, please immediately notify the sender via telephone or return mail.
> > _______________________________________________
> > gpfsug-discuss mailing list
> > gpfsug-discuss at spectrumscale.org
> > http://gpfsug.org/mailman/listinfo/gpfsug-discuss
> >
> > ________________________________
> >
> > Note: This email is for the confidential use of the named addressee(s)
> only and may contain proprietary, confidential or privileged information.
> If you are not the intended recipient, you are hereby notified that any
> review, dissemination or copying of this email is strictly prohibited, and
> to please notify the sender immediately and destroy this email and any
> attachments. Email transmission cannot be guaranteed to be secure or
> error-free. The Company, therefore, does not make any guarantees as to the
> completeness or accuracy of this email or any attachments. This email is
> for informational purposes only and does not constitute a recommendation,
> offer, request or solicitation of any kind to buy, sell, subscribe, redeem
> or perform any type of transaction of a financial product.
> > _______________________________________________
> > gpfsug-discuss mailing list
> > gpfsug-discuss at spectrumscale.org
> > http://gpfsug.org/mailman/listinfo/gpfsug-discuss
>
>
> ________________________________
> The materials in this message are private and may contain Protected
> Healthcare Information or other information of a sensitive nature. If you
> are not the intended recipient, be advised that any unauthorized use,
> disclosure, copying or the taking of any action in reliance on the contents
> of this information is strictly prohibited. If you have received this email
> in error, please immediately notify the sender via telephone or return mail.
> _______________________________________________
> gpfsug-discuss mailing list
> gpfsug-discuss at spectrumscale.org
> http://gpfsug.org/mailman/listinfo/gpfsug-discuss
>
> ________________________________
>
> Note: This email is for the confidential use of the named addressee(s)
> only and may contain proprietary, confidential or privileged information.
> If you are not the intended recipient, you are hereby notified that any
> review, dissemination or copying of this email is strictly prohibited, and
> to please notify the sender immediately and destroy this email and any
> attachments. Email transmission cannot be guaranteed to be secure or
> error-free. The Company, therefore, does not make any guarantees as to the
> completeness or accuracy of this email or any attachments. This email is
> for informational purposes only and does not constitute a recommendation,
> offer, request or solicitation of any kind to buy, sell, subscribe, redeem
> or perform any type of transaction of a financial product.
> _______________________________________________
> gpfsug-discuss mailing list
> gpfsug-discuss at spectrumscale.org
> http://gpfsug.org/mailman/listinfo/gpfsug-discuss
>
>
>
> ------------------------------
>
> Note: This email is for the confidential use of the named addressee(s)
> only and may contain proprietary, confidential or privileged information.
> If you are not the intended recipient, you are hereby notified that any
> review, dissemination or copying of this email is strictly prohibited, and
> to please notify the sender immediately and destroy this email and any
> attachments. Email transmission cannot be guaranteed to be secure or
> error-free. The Company, therefore, does not make any guarantees as to the
> completeness or accuracy of this email or any attachments. This email is
> for informational purposes only and does not constitute a recommendation,
> offer, request or solicitation of any kind to buy, sell, subscribe, redeem
> or perform any type of transaction of a financial product.
> _______________________________________________
> gpfsug-discuss mailing list
> gpfsug-discuss at spectrumscale.org
> http://gpfsug.org/mailman/listinfo/gpfsug-discuss
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gpfsug.org/pipermail/gpfsug-discuss_gpfsug.org/attachments/20170324/d200b3e3/attachment.htm>