On Mon, Feb 22, 2016 at 8:29 AM, Drew Green, MCSE, CEH <[email protected] > wrote:
> Thanks for the suggestion. According to the man page the version I'm > running uses 128 threads by default, but offers a "Threads" global > directive to adjust this. I just changed this to 512, so we'll see how > that affects it. I've been getting notifications all morning about the > service restarting, so if we can go an hour without one I think we're in > good shape. I'll let you all know if that was it. > Just providing my two cents that this also sounds to me like a Threads issue. Starting with v2.6, thread went from being allocated dynamically to a static pool. So if you were running some variant of 2.5 or lower and then upgraded to the latest (2.7 whatever), it is likely a Threads problem, as the default is pretty low. > Thanks! > > -Drew > > Drew Green, MCSE, CEH > Director of Information Technology & Security Services > <https://www.tjtsec.com> > <http://www.tjtpa.com> > 4700 Falls of Neuse Road > Suite 400 > Raleigh, NC 27609 > > Direct: 919-532-7925 > Main: 919-571-7055 > Fax: 919-516-0277 > Email Disclaimer <http://www.tjtpa.com/email-disclaimer> > > -----Original Message----- > From: Bussi Andrea [mailto:[email protected] <[email protected]>] > Sent: Monday, February 22, 2016 4:21 AM > To: [email protected] > Cc: Drew Green, MCSE, CEH <[email protected]> > Subject: Re: [Pound Mailing List] Pound hanging constantly > > Did you look at your Threads setting too? > > I'm not sure abount Pound 2.7, but I think the 2.6 version uses a pool of > a fixed number of threads, based on your configuration; the default value > could be too low for your increased load. > > Andrea > > > On 02/20/2016 03:05 PM, Drew Green, MCSE, CEH wrote: > > For the past month or two I've been dealing with an issue where pound > > will randomly hang for several minutes, and then eventually correct > > itself. Manually restarting the process will also correct this. I > > think it has to do with the load on the server. We increased the load > > in the time frame mentioned, though not by anything unreasonable. I > > spun up a new VM and tried both the custom PCI version from github as > > well as the latest stable 2.7 release, but both have the same issue. > > We were running several sites plus Exchange OWA/ActiveSync/Outlook > > Anywhere on the system, but as of Thursday are only running Exchange > > services, and the problem persists. Exchange services are where the > > increased load came from as we are now running 20 or so laptops > > through Outlook Anywhere constantly, as opposed to maybe half that > > before. I tried increasing the number of files open allowed, per a > > mailing list thread from some time back, but it's still having the > > issue. I am monitoring it with a script I wrote to check for > > availability, which will email me and restart the service if it's > > down. Everything was fine from last evening until around > > 8 this morning, when people started working again, so it seems it's > > connected to the Outlook Anywhere connections, since that's what would > > have started around that time this morning. ActiveSync runs 24/7, so > > if that was the problem it shouldn't have stopped hanging during the > > off-hours. I don't see anything in the logs that appears to be > > helpful, so I'm hoping someone here can suggest a debug log setting or > > a fix, or anything. Log level is currently set to 2. Is there a way to > > get more information than that? > > > > Thanks, > > > > Drew Green > > > > > > Drew Green, MCSE, CEH > > Director of Information Technology & Security Services > > <https://www.tjtsec.com> <http://www.tjtpa.com> > > 4700 Falls of Neuse Road > > Suite 400 > > Raleigh, NC 27609 > > > > Direct: 919-532-7925 > > Main: 919-571-7055 > > Fax: 919-516-0277 > > Email Disclaimer <http://www.tjtpa.com/email-disclaimer> > > > > -- > To unsubscribe send an email with subject unsubscribe to [email protected]. > Please contact [email protected] for questions. >
