[Glass] Nginx , FastCGI and GLASS problem

Dale K. Henrichs dale.henrichs at gemtalksystems.com
Fri Dec 6 08:50:29 PST 2013


I'll gen a 2.4 script up later today ... it is  possible that on a loaded system, the sigabort gets sent ... not responded to ... stone shrugs and sends lost ot ... gem wakes up aborts, starts handling request, receive lost OT ... 

I haven't analyzed statmons for this kind of issue before, but do you see a stat for when the sigabort was sent from the stone? There might be something in the stone log as well ...


----- Original Message -----
| From: "Johan Brichau" <johan at yesplan.be>
| To: "Dale K. Henrichs" <dale.henrichs at gemtalksystems.com>
| Cc: glass at lists.gemtalksystems.com, "Mariano Martinez Peck" <marianopeck at gmail.com>
| Sent: Friday, December 6, 2013 8:10:01 AM
| Subject: Re: [Glass] Nginx , FastCGI and GLASS problem
| On 06 Dec 2013, at 02:30, Dale K. Henrichs
| <dale.henrichs at gemtalksystems.com> wrote:
| > If this is happening in 2.4 (Johan), I'll have to do a little more
| > research to see if there is an exception equivalent to
| > RepositoryViewLost signalled.
| Yes, those things happen in in our "busy" repositories, which are on
| 2.4
| If there is an equivalent, I would be glad to install that handler.
| I am indeed also suspicious if it occurs on a loaded system. Though,
| statmonitor has learned me that the Tx state of a gem was 'in
| transaction' even when this 'termination from stone' occurs.
| Will try to retrieve that statmonit file again...
| Johan

More information about the Glass mailing list